S P E C I F I C A T I O N
HEALTHCARE DATA INTERCHANGE SYSTEM AND METHOD
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority to United States Provisional Patent Application Serial
No. 62/104,532, filed on January 16, 2015, the disclosure of which is expressly incorporated
herein by reference in its entirety and for all purposes.
BACKGROUND
[0002] Many countries and states within the United States are facing an impending
healthcare crisis due to declining population health, increasing prevalence of chronic diseases,
and the extraordinarily high cost of caring for patients in an acute care system. Current disease
management programs rely primarily on manual, siloed interventions, which are labor intensive
and un-scalable.
SUMMARY
[0003] The present disclosure relates to a health data system that is part of a scalable technology
core that can be integrated into local healthcare infrastructure to create a care management
framework for delivering patient-centric and value-based care in a community, setting the stage
for scalability to targeted communities.
[0004] In accordance with a first aspect disclosed herein, there is set forth a health data
system for delivering patient-centric and value-based care, comprising:
[0005] a health data server;
[0006] one or more health data sources in communication with the health data server over a
secured network, wherein said health data sources each have a set of polling permissions;
[0007] one or more agent modules of the health data server that poll health data from the data
sources at a designated frequency based on a set of identifiers and the set of polling permissions;
[0008] a first switch module for providing the polled health data into a common information
model, the common information model being defined by at least one patient record, each patient
record having one or more attributes; and
[0009] one or more interface modules for gaining access to the common information model
based on a set of access permissions.
[0010] In some embodiments of the disclosed system, the common information model
includes a distributed database and wherein the one or more attributes optionally define at least
one of clinical health data, laboratory data, remote monitoring data, biometrics, wearables, social
media data, self-reported data, mobile application data, and device instrumentation.
[0011] In some embodiments of the disclosed system, the first switch module further
augments the common information model with a new attribute when the polled health data does
not map into the one or more attributes.
[0012] In some embodiments of the disclosed system, the first switch module at least one of
filters the polled health data prior to providing the polled health data into the common
information model based on storage permissions, the storage permissions optionally being
provided by at least one of the health data server and the one or more health data sources,
matches patient records against each other, and controls network connections over the secured
network.
[0013] In some embodiments of the disclosed system, the designated frequency is set by at
least one of the health data server and the one or more health data sources.
[0014] In some embodiments of the disclosed system, the system includes a second switch
module for providing the polled health data into the common information model, wherein said
first switch module communicates with said second switch module for receiving the polled
health data.
[0015] In accordance with another aspect disclosed herein, there is set forth a method for
delivering patient-centric and value-based care, the method comprising:
[0016] polling one or more health data sources for health data via one or more agent modules
of a health data server, wherein each of the one or more health data sources have a set of polling
permissions;
[0017] populating a common information model with the polled health data via a first switch
module, the common information model being defined by at least one patient record, each patient
record having one or more attributes; and
[0018] providing access to the common information model via one or more interface
modules based on a set of access permissions.
[0019] In some embodiments of the disclosed method, populating the common information
model comprises populating a distributed database optionally with the one or more attributes
selected from at least one of clinical health data, laboratory data, remote monitoring data,
biometrics, wearables, social media data, self-reported data, mobile application data, and device
instrumentation.
[0020] In some embodiments of the disclosed method, the method further comprises
augmenting the common information model with a new attribute when the polled health data
does not map into the one or more attributes via the first switch module.
[0021] In some embodiments of the disclosed method, the method further comprises filtering
the polled health data based on a set of storage permissions prior to said populating.
[0022] In some embodiments of the disclosed method, the set of storage permissions are
provided by at least one of the health data server and the one or more health data sources.
[0023] In some embodiments of the disclosed method, polling occurs at a designated
frequency set by at least one of the health data server and the one or more health data sources.
[0024] In some embodiments of the disclosed method, the method further comprises
matching patient records via the first switch module.
[0025] In some embodiments of the disclosed method, the method further comprises polling
a second switch module via the first switch module for the polled health data.
[0026] In some embodiments of the disclosed method, polling is limited by the set of polling
permissions.
BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Fig. 1 is an exemplary network diagram illustrating an embodiment of a health data
system.
[0028] Fig. 2 is an exemplary network diagram illustrating an alternative embodiment of the
health data system of Fig 1.
[0029] Fig. 3 is an exemplary flow chart illustrating an embodiment of a method of obtaining
and processing healthcare data by the health data system of Figs. 1 and 2 .
[0030] Fig. 4 is an exemplary flow chart illustrating an embodiment of a method of
processing and providing healthcare data by the health data system of Figs. 1 and 2 .
[0031] Fig. 5 is an exemplary method of defining a consumer health problem, creating an
ecosystem to resolve it, monitoring the care delivery network to identify where health benefits
are realized, and isolating the associated cost savings in order to redistribute them to various
stakeholders.
[0032] Fig. 6 is an exemplary network diagram illustrating an embodiment of health care
ecosystem.
[0033] It should be noted that the figures are not drawn to scale and that elements of similar
structures or functions are generally represented by like reference numerals for illustrative
purposes throughout the figures. It also should be noted that the figures are only intended to
facilitate the description of the preferred embodiments. The figures do not illustrate every aspect
of the described embodiments and do not limit the scope of the present disclosure.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0034] Emerging technology solutions are poised to transform health care delivery. The
health data system 100 as illustrated in Figs. 1 and 2 can be used along with other systems to
implement a suite of technology-enabled data-driven solutions designed to augment and
accelerate effective disease management and care. For example, the health data system 100 can
be part of a scalable technology core that can be integrated into local healthcare infrastructure to
create a care management framework for delivering patient-centric and value-based care in a
community, setting the stage for scaling to a broader set of communities.
[0035] In various embodiments it can be beneficial to configure a system that is consumercentric
with mobility and access, self-service enabled and designed with data fluidity and
continuity. It can also be beneficial to have a personalized system that is configured to treat a
person, not a diagnosis (e.g., by using a contextualized health profile and by leveraging
longitudinal physiological plus behavioral, social and environmental data). It can be further
beneficial to have an outcome-based system wherein value is driven by health outcomes; where
quality is defined by safe and evidence-based care; where efficiency is achieved through
optimized allocation of capacity, capability, availability and cost; and where effectiveness is
personalized based on personal preference and ability, impacted by social and environmental
factors. A desirable system can also include proactive health management that extends beyond
reactive episodic care, includes population segmentation and stratification, includes a chronic
disease care coordination plan, includes a long term health management plan, and includes a
consumer education plan.
[0036] A health data system 100 that provides for obtaining, storing, curating, analyzing
and/or providing access to health data can prove desirable and provide a basis for a wide range of
applications as described in detail herein. This result can be achieved, according to one
embodiment disclosed herein, by a health data system 100 as illustrated in Fig. 1.
[0037] Turning to Fig. 1, the health data system 100 is shown as comprising a plurality of
data-source devices 110, a health data server 120, and a plurality of user devices 130. In some
embodiments, the user devices 130 represent user services available from the health data system
100. The data-source devices 110 and user devices 130 are connected to the health data server
120 by a secure network 140 that can consist of any combination of wireless and wireline links.
The data-source devices 110 are shown as comprising a smart phone data-source 110A, a laptop
data-source HOB and a server data-source 1IOC, but in further embodiments, any suitable device
can comprise a data-source 110, including a desktop computer, a tablet computer, a gaming
device, a smart-television, a headset computer, a smartwatch, a body monitor device, or the like.
Additionally, various embodiments can include any suitable number of any such data-source
devices 110.
[0038] Similarly, although the user devices 130 are shown as being invoked from a smart
phone user device 130A and a laptop user device 130B, in further embodiments, a user device
130 can comprise any suitable device including a server, a desktop computer, a tablet computer,
a gaming device, a smart-television, a headset computer, a smartwatch, a body monitor device,
or the like. Additionally, various embodiments can include any suitable plurality of any such
user devices 130.
[0039] The server 120 can include one or more server systems, which can include any
suitable plurality of devices and/or a cloud-based system. Additionally, the server 120 can
comprise a plurality of modules, databases, or the like. For example, Fig. 2 illustrates one
embodiment of the health data system 100 that comprises one or more agent modules 205, a
security module 309, a market rules module and/or database 310, a switch module 311, a
common info model module and/or database 312, a big data store 313 and an API 314, which are
part of a cloud-based server system 120. The cloud-based server system 120 shown in Fig. 2 is
implemented as a private cloud for illustration purposes only.
[0040] In accordance with various embodiments, the server 120 is configured to receive,
process and store data obtained from data source devices 110 (see, e.g., Fig. 3). The server 120
can also be configured to process and/or retrieve stored data and provide it to the one or more
user devices 130 in response to various queries or data requests that the user devices 130 may
provide (see, e.g., Fig. 4).
[0041] Referring to Figs. 1 and/or 2, one or more data source device 110 can be associated
with one or more data source 201. For example, data sources 201 can include a variety of
potential stakeholders and their associated data sources that might approve data for sharing via
the health data system 100 in a targeted health ecosystem. For example, data sources 201 can
include healthcare providers (e.g., data can be electronic medical records, lab data), health
insurers / payers (e.g., data can be claims records), pharmaceutical and medical device
companies (e.g., data can be clinical trial records, adverse event data), research (e.g., data can be
a genomic profile), government/community health programs (e.g., data can be population health
statistics) partnership databases), and/or individual patients (e.g., data can be biometrics,
activity/behavior). Each data source 201 can be independently owned with its own set of unique
data access control rules. Accordingly, the health data system 100 advantageously provides
access to disparately held data (e.g., across the data sources 201) for use by independently
developed health data services (e.g., via the user devices 130).
[0042] Each stakeholder data source 201 can choose the specific fields and elements, or
subsets of data, which they approve to share, and the system 100 can manage the approvals of
identified data (e.g., through consent/data use agreement) and/or identifiable data (e.g., through
consent/ Business Associate Agreements) as described in more detail herein and as illustrated in
Fig. 3 . Although examples of data associated with a given data source are described above, data
sources can provide or be the source of any suitable type of data without limitation.
[0043] Agents A-C 306, 307, and 308 of Fig. 2 represent an example architecture of the
distributed agent modules 205 that are configured to obtain, receive and/or access data on a
manual and/or automated basis from the data sources 201. The agent modules 205 can be
associated with the one or more data sources 201 and/or the data source devices 110, and a given
data source 201 or data source device 110 can be associated with one or more agent modules
205. These agent modules 205 ensure that any needed metadata / supporting elements (e.g.,
consents, access rights, source information) are transmitted alongside data that is obtained,
received and/or accessed from the data sources 201.
[0044] In some embodiments, manual interactions can be conducted via a web-enabled portal
within the data source 201 (e.g., via one or more data source device 110), in which an owner of
the data source 201 is responsible for deciding what data is transmitted to the server 120 from the
data source 201 by agent modules 205. Automated interactions can be via script programs that
are configured (e.g., with business rules, or the like) and then scheduled to run on a particular
frequency, and/or in real time based on desired monitoring criteria to specify what data is
transmitted to the server 120 from the data source 201 by the agent modules 205. The data in the
data source 201 that can be transmitted to the server 120 by the agent modules 205 is designated
by an identifier. An identifier is a string of alpha-numeric characters that uniquely identify a
patient record in the data source 201.
[0045] Business rules can be configured in the agent modules 205 and can include naming
conventions, data lineage tracking, permissible and non-permissible fields based on the ability to
share identifiable data, sharing restrictions, and other stakeholder organization-specific rules for
the agent to follow when managing data. In some embodiments, as illustrated in Fig. 2, agent
modules 205 can encrypt data that will be transmitted to the server 120 prior to leaving
environment and/or firewall of data source 201.
[0046] In various embodiments, data that is not in accordance with the sharing business rules
defined by the data source 201 will remain in the environment of the data source 201 and will not
be brought into the health data system 100 by the agent modules 205. In some embodiments,
this can include identifiable data that does not contain a flag indicating that a patient's consent
for sharing was obtained and a notice of privacy practices was given, such that the health data
system 100 must infer that the data source 201 does not have permission from an individual to
share the individual's data outside of the environment of the data source 201 (e.g., blocks 403,
406 and 409 of Fig. 3).
[0047] In one embodiment, the implementation of a selected agent module 205 includes the
following steps:
1. The agent module 205 is designated with the IP address of the data device 110
that holds the data provided by a data source 201.
2 . The agent module 205 is designated with identifiers for patient records available
from the data source 201. The identifiers are provided by the data source 201 to
the health data server 120.
3 . The agent module 205 is designated with a frequency for obtaining data on patient
records from the data source 201 (using the identifiers set in step 2). The
frequency may be set by the health data server 120 and/or the health data sources
201.
4 . The secure network 140 establishes a dedicated connection between the health
data system 100 and the data source 201.
5 . The agent module 205 collects data from the data source 201 on the identifiers
(set in step 2), from the location (set in step 1) per designated frequency (set in
step 3) using the dedicated network connection (set in step 4).
6 . Data collected by the agent module 205 is stored in a message queue in the health
data server 120. This message queue is processed by the switch 3 11. The agent
module 205 represents the data collected from the data source as a list of
(, ) pairs.
7 . The agent module 205 continues to run until the health data server 120 is notified
by the data source 201 to no longer collect data from the data source 201.
8 . At any time the data source 201 can modify the list of identifiers on whom data
can be collected by the agent module 205 for transmission to the server 120.
9 . At any time the data source 201 can send a message to the health data server 120
to terminate data available from the data sources 201 in the health data server 120
on one or more (or all) identifiers on whom the agent module 205 may have
collected data from the data source 201 (in step 5).
10. At any time the data source can obtain from the health data server 120 a report on
what data was collected by agent module 205 from the data source 201.
[0048] As shown in Fig. 2, the server 120 can comprise a security module 309, which can
comprise one or more security components that are operable to ensure communications to and
from the server 120 are authorized/authenticated and/or encrypted in accordance with data source
stakeholder needs, regulatory requirements, and industry best practices (see, e.g., blocks 403,
406 and 409 of Fig. 3). For example, the security module 309 and components thereof can be
configured to ensure that any agent 205 or direct user interactions with the health data server 120
are in compliance with applicable business rules or the like. In some embodiments, the security
module 309 can comprise physical and/or virtual security components, which can include the
following components, or the like.
Security Incident and Event Management:
• Log Collection, Correlation and Notification
• External Internet Attack & Threat Monitoring
• Internal Attack & Threat Monitoring
Identity & Access Management:
• API Gateway Firewall
• Identity Access Management (IAM)
• Advanced Multi Factor Authentication
• Fine Grained Entitlement Access Control
• Privileged Access Management
Database Security:
• Masking of database information from administrators
• Database Auditing and Accounting of Access
• Database Encryption
Network Security:
• Unified Threat Management (UTM) Firewalls
• Data Leak Prevention
• Configuration Management & Monitoring
• Vulnerability Management Scanning
• 3rd Party Digital Certificates
• Attack & Penetration Tests
• Load Balancer
End Point Security:
• Hypervisor VM Firewall Security
• Encryption Vault Key Protection
• Patch Management Protection
• Malware Protection
[0049] The health data server 120 can include a market rules module 310 that can control a
portion of or all data flow that occurs within the system 100 (see, e.g., blocks 4 11 and 414 of Fig.
3). A rule in the market rules module 310 specifies access to some data in server 120 from data
source 201 by the user device 130.
[0050] In one embodiment, market rules module 310 is characterized by the following steps:
1. To configure the market rules module 310 with data access rules, called market
rules, for each data source 201 connected to the health data system 100, set up a
configuration file containing rules provided to health data server 120 by data
source 201
2 . A market rule has the following specification:
a . ( ): this rule states that user
device 130 identified by can have access to a record, identified
by , provided by data source 201, identified by , and
available for access by a user of the user device 130, identified by .
3 . To validate a request for data by user device 130 the market rules module 310
checks if the specific record requested by the user device 130 from data source
201 is allowed by the market rules specified by data source 201. A market rule
can be specified to inform the switch 311 to not persist data from 201 collected by
205 in the common information model 312. This market rule can have the
following specification: ( on-demand): this rule states that
a record identified by from the data source 201 identified by should not be stored in the common information model 312. For records from
the data source 201 identified by this rule, the switch module 3 11 stores a query in
the common information model 312 for this record and executes this query at the
time a request is made for this record by the API 314.
[0051] In various embodiments, business and/or market rules can facilitate the application of
algorithms and logic such as data consistency checking and cleansing, reference data
standardization, and master patient indexing, in order to facilitate reuse and avoid duplication or
mis-representation of data. In various embodiments, this can include operations such as
identification and adjusting of null values and inconsistencies in units of measure, and the usage
of demographic values to align multiple stakeholders' records for the same individual under a
single global identifier within the system 100. Accordingly, in some embodiments, business
and/or market rules can enable the use of data across a plurality of services to derive novel
insights while also protecting information privacy.
[0052] The health data server 120 can also include one or more switch modules 3 11. In one
embodiment the implementation of a switch module 3 11 includes:
A selected switch module 311 is set up with an input message queue. The input
message queue receives messages from the agent module 205. The format of
each message can include a list of (, ) pairs.
The selected switch module 311 is set up with a data quality queue to log
messages identifying records obtained from the data source 201 that have been
found to have data quality issues by the data quality engine (DQE). DQE
includes:
a . Data quality rules for each data source 201 connecting into system 100
b. Data quality rule can include the following format:

c . Application of data quality rules corresponding to the data source 201 to
messages placed into the input message queue of the switch module 311 by
the agent module 205 to flag a data quality issue if any attribute of the
message in the input message queue does not have a value contained in the set
of possible values for that attribute in the data quality rules for that data source
201.
The health data server 120 sends to the data source 201 data quality messages
from the data quality queue for resolution by the data source 201
If no data quality issue is detected with the message in the input message queue
then the switch module 3 11 loads up the data in the message into the
corresponding table in the common information model 312. For example, if the
message is about the medication administered to a given patient at a hospital then
the contents of this message are loaded into the medication table in common
information model 312 tagged with the identifier for that patient along-with the
identifier of the data source 201 associated with the hospital from where the agent
module 205 collected the data for that patient. At the time of loading the data into
the common information model 312, one or more coding dictionaries can be used
to map terms in the message into standard terms. For example, blood pressure
can be mapped to hypertension.
The switch module 3 11 periodically executes the match engine. Match engine
detects when data from two different data sources 201 belongs to the same
patient.
The switch module 311 is set up with an output message queue. Output message
queue receives messages from the API module 314.
The switch module 3 11 invokes the market rules module 310 for each message in
its output message queue. If the market rules module 310 validates the message
for access to requested data from 312 then the switch module 3 11 maps the
message in the output message queue into a query for common information model
312. The query format is (, ) where identifies
a data source 201 and identifies a record from source 201. This query
is then executed against common information model 312 by switch component
3 11. The data obtained from common information model 312 by switch 311 is
returned to the user device 130 by the API 314.
The switch module 3 11 applies a natural language processing ( LP) engine to
unstructured attributes in data from the data sources 201 collected by the agent
modules 205 (unstructured attributes are identified in 312 as attributes that allow
arbitrary length character strings as values). The resulting data includes
(, ) lists and these lists are used to add data to the
corresponding tables in the common information model 312. In this way the
switch component 3 11 reduces unstructured data from 201 into structured data in
the common information model 312. By doing so, the switch component 3 11
integrates structured data from the data sources 201 with unstructured data from
the data sources 201.
9 . The switch module 3 11 tags each data stored in the common information model
312 with any data used in the generation of the data being stored in the common
information model 312. This way complete data traceability is maintained in the
common information model 312. For example, if there two sets of medication
records on a patient in 312 and a user service 130 generates a reconciled
medication list then prior to storing the reconciled medication list in the common
information model 312, the switch module 3 11 tags the reconciled medication list
with the identifiers for the two medication list that have been reconciled.
10. The switch module 3 11 implements a data exclusion service to de-identify data
from the data sources 201. This service is provided with a list attributes that need
to be de-identified. Upon execution of the data exclusion service on specified
data the values of the identified attributes are masked to make it impossible to
obtain the original values of these attributes by replacing each byte of storage
allowed for holding a value for an attribute with the null byte '' .
Simultaneously, the data is tagged with the names of attributes that are masked. In
this way the input data is rendered de-identified by the data exclusion service. In
an alternative embodiment of data exclusion service the value of a de-identified
attribute can be replaced with a unique tag and the correspondence between the
original value of the attribute and the replacement tag can be added to a
tokenization table in the common information model 312.
[0053] The switch component 3 11 can implement various services to manage the lifecycle of
data from onboarding to termination, including:
• Registration service to onboard & register the data source 201 and the data service 130.
The registration of the data source 201 with system 100 requires the data source 201 to
provide an interface to system 100. This interface consists of an IP address and a set of
instructions on how to obtain data from that IP address. The data source 201 provides a
set of data access rules to health data server 120 that govern the use of the data provided
by the data source 201 by the user devices 130 participating in system 100. In health data
server 120, these rules are called market rules configured in the market rules module 310.
Registration by data service with the system 100 requires system 100 to provide to the
data service 130 an IP address and a set of instructions on how to obtain data from that IP
address. Before responding to a request for data by a registered data service 130 system
100 validates the request against market rules provided by data sources 130 who have
contributed requested data into system 100. This way access to data in system 100 is
always controlled by data access rules established by the data sources 201 contributing
data into system 100.
Data audit service to log data entering and leaving the system 100.
Data quality engine (DQE) to monitor the quality of data entering the system 100 which
in turn drives better analytics, which can include the following dimensions: Attributelevel
data quality (for example age should be in a certain range or last name should not be
blank), Context or aggregate data quality (for example patient data on average should be
approximately 50% male 50 % female plus or minus certain margin), Operational data
quality (for example, reject if patient demographics data is present but medication data is
not present thereby rendering the entire patient record not useful)
Data termination service to remove from the system 100 any data from a data source
upon request by the data source owner makes it easy for the source owner to remove their
data with control from the platform.
In some embodiments, where the rules cannot sufficiently resolve inconsistencies in the
data, those elements/fields are flagged for manual intervention from a data steward user
who is authorized to see identifiable data from all stakeholders and reconcile the
differences. This can include examples in which patient matching / indexing algorithms
have identified two records, from two stakeholders, which records appear to belong to the
same patient but are missing.
• Patient match service to identify and match data on the same patient ingested from
multiple data sources 201. For example, using patient record elements that are less likely
to change throughout a patient's life (e.g., name, date of birth, gender, SSN—as opposed
to a phone number or address for example), the health data server 120 can match a patient
across multiple records of different types from different sources. Accordingly, in some
embodiments, the health data server 120 can be configured to assign a weight to each
field to create a total match score, with the weight implying the importance of that field
in the matching process (e.g., social security number (SSN) and date of birth (DOB) can
get a 35% weight while gender, last name, and first name can get 10% weight each),
which weighting may be operable to improve accuracy of the match. Upon matching
data from two or more data sources 201, the switch module 3 11 tags the data with a
unique id for that patient in the health data system 100.
[0054] The resulting cleansed, structured, standardized data can be stored in one or more
databases defined by the common information model module 312, (see, e.g., block 413 and 415
of Fig. 3). For example, the common information model module 312 can standardize
information in accordance with terminologically robust standards such as Systematized
Nomenclature of Medicine—Clinical Terms (SNOMED) standards (e.g., for procedure codes,
medication method), International Organization for Standardization (ISO) standards (e.g., for
Country Codes), Fast Healthcare Interoperability Resources (FHIR) standards (for certain fields),
or the like. For example, medication route and dosage method can be standardized to the
SNOMED values, country code can be standardized to 3 digit ISO-3166 country codes, gender
and race can be standardized using FHIR standard values, or the like.
[0055] In some embodiments, when standardizing and integrating data, the system 100 can
also facilitate patient matching. For example, using patient record elements that are less likely to
change throughout a patient's life (e.g., name, date of birth, gender, SSN—as opposed to a phone
number or address for example), the system 100 can match a patient across multiple records of
different types from different sources. Accordingly, in some embodiments, the system 100 can
be configured to assign a weight to each field to create a total match score, with the weight
implying the importance of that field in the matching process (e.g., social security number (SSN)
and date of birth (DOB) can get a 35% weight while gender, last name, and first name can get
10% weight each), which weighting may be operable to improve accuracy of the match.
Additionally, the business/market rules can also identify unstructured data and prepare it for
storage in clusters separate from the rest of the structured data (e.g., in big data storage 313)
without losing lineage information (see, e.g., block 412 of Fig. 3).
[0056] Although various embodiments discussed herein relate to processing data for storage
on the health data server 120, in further embodiments, data processing described herein can
apply to data in-motion. In other words, in some embodiments, data may not be stored on the
server 120 and can be passed between data sources 201 and data consumers 207 using the health
data server 120 as an intermediary, but without the data being stored in the health data server
120. In such embodiments, data can be processed as described herein.
[0057] Referring to Figs. 1, 2, and 4, the system 100 can comprise a plurality of data
consumers 207 that are configured to consume data from the server 120 for the purpose of
research, clinical care, commercial purposes, or the like, using the one or more user devices 130.
For example, a given data consumer 207 can comprise one or more user devices 130 that are
configured to request and/or receive various types of data from the server 120 as described in
more detail herein and as shown in Fig. 4, through the user devices 130.
[0058] In various embodiments, data consumers 207 can include healthcare providers (e.g.,
physician from a hospital sharing of electronic medical records with a primary care physician for
continuity of care, or the like), health insurers/payers/risk-bearing entities (e.g., using outcomes
data to adjust formulary status for pharmaceuticals), pharmaceutical and/or medical device
companies (e.g., using outcomes data to support post-marketing surveillance evaluations / Stage
IV clinical trials), government/community health programs (e.g., using provider data to support
at-home care, especially as it relates to environmental factors), and/or individual patients (e.g.,
medication compliance reminders), who have agreed protect data and use it in accordance with a
data use agreement and/or business associate agreement.
[0059] Although Fig. 2 shows data sources 201 being separate from data consumers 207, in
some embodiments a data source 201 and data consumer 207 can be the same. Additionally, a
data source device 110 can also include a user device 130, and vice versa.
[0060] In various embodiments, data consumers 207 can request data either manually (e.g.,
via mobile application request initiated by a provider, patient, advocate, or the like) or via
automated requests (e.g., automated pull from the health data server 120 common information
model database 312 into a local research database (not shown) on a daily, weekly, monthly basis,
or the like).
[0061] In various embodiments, the application programming interface (API) 314 can be
configured to handle requests for data from the data consumers 207 according to a published
catalog of service calls (e.g., requests for patient demographics, or queries for various subsets of
data). Such interactions can require authorization and authentication, as facilitated by the
security components 309. In some embodiments, requests from the API 314 can be managed by
the switch 3 11 according to the market/business rules module 310. Applicable rules can grant or
block access for certain requests, based on the level of access associated with each request.
[0062] For example, a request for identifiable data by the user device 130 can be blocked if
the requesting user (or user's organization/role type) has not been granted rights by the original
data source 201 or patient who shared that data (e.g., as configured in the agent 205 that handled
the data). In another example, rules can also specify that only an organization's own users can
see identifiable data, while all others may only receive access to identified versions or subsets of
the organization's data (see, e.g., block 508 in Figure 4). Depending on the nature of the access
restriction, the user may receive an error message (see, e.g., block 519 in Figure 4).
[0063] In some embodiments, the server 120 can determine whether a request is granted by
assessing each request based on three checks:
• Regulatory Compliance: does the request follow Health Insurance Portability and
Accountability Act (HIPAA) standards and/or Health Information Trust Alliance
(HITRUST) regulations. In other words the health data server 120 checks that proper
regulatory documentation is in place (e.g., a HIPAA business associate agreement
(BAA)).
• Security: the health data server 120 ensures authentication and authorization.
• Market Rules: the health data server 120 implements organization-specific controls that
govern data access rules.
[0064] For example, compliant requests/data output, serviced by the API module 314, can
include a user requesting access to data for which he or she is an authorized user; the data source
201 requesting that their data on the health data server 120 platform is terminated; a patient
requesting that their data not be shared; a patient requesting that certain specific parties only can
see their data; a Fast Healthcare Interoperability Resources (FHIR) data output of patient
demographics data for a Medical Record Number (MRN) for which the requestor has been
granted; a request for Tableau (Tableau, Inc., Seattle WA) visualization of age demographics of
multiple patients being viewed by a user who is authorized to access these multiple patients; a
request for analytics comparing a patient to other "patients like me" across organizations through
a third party clinical decision support solutions; and the like.
[0065] Examples of non-compliant requests, serviced by the API module 314, (for which no
data output is provided) can include: an individual or organization requesting access to data for
which they are not supposed to have access; a request to bypass security policies and procedures
(requesting to share user id for example); a request to disable or change a market rule from an
unauthorized individual; request to be able to direct Structured Query Language (SQL) query
against health data server 120 data stores; a request to use information in a way that falls outside
the scope of the data use agreement; a business entity requesting access to protected health
information on a patient without a HIPAA business associate agreement (BAA); and the like.
[0066] However, if permissions are granted, then the applicable query can be run against the
common information model 312 {see, e.g., blocks 509, 510, 5 11 in Figure 4).
[0067] As an example, the output of the common informational query might be a data set in
relational format, while the output of the big data query might be a frequency and sentiment
analysis for a bolus of unstructured content and specific keywords therein.
[0068] The result of the query or queries can be consolidated into a unified output (see e.g.,
block 512 in Figure 3), to the extent it is not already aggregated, that can be reflective of the
combined data of various contributing data source 201 stakeholders, as filtered or defined by the
user, alongside the necessary metadata that is required to interpret the data. For example, in one
embodiment, the combined data from a provider, a payer, and a community health program can
be aggregated into a total outcomes data set, matched against a set of reference data, and output
in the common information model 312. This data set can be further checked to ensure that the
requesting user and/or application has sufficient access privileges (see, e.g., block 513 in Figure
4) and has the appropriate contracts / data use agreements / business associate agreements in
place, if needed.
[0069] Following this, the data can sent out according to services calls of the API 314 (see,
e.g., block 514 in Figure 4) in encrypted format (see, e.g., blocks 515 and 516 in Figure 4) to the
requesting users / services. The users can now (see, e.g., blocks 517 and 518 in Figure 4) make
use of a broader set of data to identify insights and make decisions. For example, a payer user
who did not previously have access to outcomes-related data other than what was present on
claims submitted may now have deeper insights into a patient's clinical experience based on a
combination of data gathered during and after the experience (e.g., in community support).
Example Use and Implementation of the Health Data System 100
[0070] Changes in the New Health Economy are forcing healthcare organizations to
collaborate in order to deliver the comprehensive care that patients need to achieve target health
outcomes. Once the health ecosystem necessary to address the needs of a patient population is
proposed, the health data system 100 can act as an information interchange that facilities
data/information exchange between members (e.g., data sources 201 and/or data consumers 207
as shown in Fig. 2). The health data system 100 can be an ecosystem enabler that identifies and
connects the right network of health organizations (e.g., data sources 201 and/or data consumers
207 as shown in Fig. 2) to improve consumer health among a targeted population. The health
data system 100 can enable third-parties (e.g., data sources 201 and/or data consumers 207 as
shown in Fig. 2) to transact in order to provide and/or consume the data and analytics necessary
to achieve various desirable business objectives. By acting as an intelligent integration engine,
the health data system 100 can facilitate connectivity across a marketplace to enable fluid
exchange of data and services.
[0071] For example, the health data system 100 can identify and enable an ecosystem of
individuals, organizations, and data-driven solutions in a program to improve disease
management. In such an example use and implementation, the health data system 100 can
connect such players as a mobile platform and a cognitive analytic solution to deliver continuous
and seamless information to appropriate stakeholders (e.g., data sources 201 and/or data
consumers 207 as shown in Fig. 2). Fig. 5 illustrates a novel method 500 to define a consumer
health problem, create an ecosystem to resolve it, monitor the care delivery network to identify
where health benefits are realized, and isolate the associated cost savings in order to redistribute
them to both the risk-bearing entities and other ecosystem members. Specifically, the method
500 comprises defining targeted consumer/patient health challenges 510; identifying ecosystem
participants and interactions required to address patient needs 520; using the health data
system/information exchange 100 to facilitate ecosystem connectivity 530; optimizing
interactions and implementing governance 540; and capturing value through consumer health
improvements to sustain ecosystem through shared cost savings 550.
[0072] The health data system 100 can be configured to be and/or be a portion of a system
that enables a comprehensive set of business and technology services, including the capability to
intelligently locate distributed data, disseminate timely insights, and support a broad community
of users across the care delivery network.
[0073] Such a health data system 100 can facilitate ecosystem interactions and act as a
trusted broker of information exchange across stakeholders; connect to a variety of analytic
applications that are able to access targeted data in a secure fashion in order to deliver intended
outcomes such as improved education, health monitoring, improved health decision making, and
disease prevention; provide access to data to facilitate ongoing reporting on outcomes of a
disease treatment and prevention program including patient outcomes, care delivery
effectiveness and efficiency, as well as value captured and savings achieved. In various
embodiments, such a system can enable such reporting by providing access to the data through
predefined dashboards, reports, score cards as well as ad-hoc analysis for a select number of
stakeholders all based on appropriate permissions and agreements. Such reporting can also
support tracking of funds flows. Additionally, such a system can be configured to ensure privacy
and security of patient and stakeholder information through a defense-in-depth security model
and develop market rules to manage interactions of members in the ecosystem.
[0074] To leverage the health data system 100 to improve care management (and other
health care programs in other embodiments) a range of health stakeholders (e.g., data sources
201 and/or data consumers 207 as shown in Fig. 2) can be connected to share information. As
discussed herein, the health data system 100 can be designed to provide access to data and
analytics from across organizations in order to drive novel insights and care transformation.
[0075] Additionally, the health data system 100 engages ecosystem participants including,
but not limited to, care delivery network entities (e.g., core providers, specialists, out-patient,
hospital, clinics, urgent care, retail clinics that provide patient care); health management network
(HMN) entities (e.g., risk-bearing entities (RBEs) such as government organizations, self-insured
employers, and commercial payers; and non-RBEs that include employers (not self-insured),
community centers, and patient advocates, which act as health promoters, educators, and the
like). In some embodiments, such HMN groups can fund, subsidize, or otherwise incentivize
target programs to optimize care delivery and reduce costs.
[0076] The health data system 100 can also be configured to engage commercial business
network entities including, but not limited to, key business entities that should have a vested
interest in driving treatment adherence and compliance given the impact of health spend on the
American economy. Such entities can also serve as coordinators to drive health management by
tailoring their products or delivery to the population and financially supporting community
health initiatives.
[0077] Additionally, the health data system 100 can integrate fit-for-purpose technology
solutions that meet the needs of the community, including a mobile patient engagement platform
or a cognitive analytics expert system.
[0078] Since currently-available healthcare ecosystems are deficient, a healthcare ecosystem
that provides for obtaining, storing and/or providing healthcare data can prove desirable and
provide a basis for a wide range of applications as described in detail herein. This result can be
achieved, according to one embodiment disclosed herein, by a healthcare ecosystem 600 as
illustrated in Fig. 6 . For example, Fig. 6 illustrates an embodiment of healthcare ecosystem 600,
which involves an exchange of information from multiple sources delivered via web services
(APIs). As shown in Fig. 6, the health data system 100 can serve as the data hub through which
various ecosystem participants can connect and share data. For example, the health data system
100 can allow stakeholders in the ecosystem 600 to can gain access to data "in motion" (e.g., the
system 100 transits data only) or "at rest" (e.g., the system 100 stores/persists the data in a
repository as discussed herein).
[0079] In various embodiments, the health data system 100 can act as an interchange, where
the system administrator does not own data or determine access rights/rules, but rather acts as a
data steward to allow data owners (e.g., patients, clients and/or stakeholders) to dictate to which
users and services the data will flow. The health data system 100 can enable the exchange of
both regulated and unregulated data, and each entity in the ecosystem 600 can provide and
receive data that is governed by relevant regulations and/or client-defined access rules (e.g.,
governed by the market rules 310 as shown in Fig. 2).
[0080] In various embodiments, the health data system 100 can enable data to be shared from
a variety of sources, including consumer applications, medical devices, Electronic Health
Records (EHRs), pharmacy records, and patient registries. Each of these data types can then
feed into a wide range of applications supported by the health data system 100.
[0081] Data types in scope for various embodiments, such as a disease prevention and
treatment program can include: provider primary clinical data (e.g., demographics, home
medications, allergies, laboratory and pathology reports, transcribed records, and the like);
provider secondary clinical data (e.g., natural language processing data, observation data,
prognosis scores, analytics and statistical data, and the like); reference data for standardization
(FHIR standard ontology, ISO 3166 codes, LOINC and SNOMED standard data sets, and the
like); patient self-managed / self-reported data (e.g., glucose meter, wearables data, and the like);
community advocate-related data (e.g., self-management program data, educational program
data, and the like); financial / risk-bearing entity data (e.g., member plans, provider quality
assessments, claims, and the like).
[0082] In various embodiments, the ecosystem 600 and/or health data system 100, can be
configured to deliver the following services:
Provides the technology platform to host data in a private, secure, scalable
cloud, including:
• Hosting and Management: Proprietary & 3 party data, analytics and
applications
• Application Development Services: Deliver APIs to enable development
of new applications
• Always On Service/ Customer Support: 7 x 24 real-time monitoring,
support, and rapid response
Core • Security: Layered security model creates safeguards to facilitate privacy
Information
and security protections and breach prevention that meets HITRUST control
Interchange
specifications
• Access Control: Serving unique information access governance and policy
requirements of ecosystem members
• Market Rules: Manage interactions of members in the ecosystem through
standard market terms and conditions, such as data access and use rules
• Monetization of Intellectual Property (Marketplace): Enable 3 parties
to transact in order to provide/consume the data and analytics necessary to
achieve business objectives
Federated Linkage of a consumer's attributes defined by and stored across heterogeneous
Identity distinct source systems
Management
Aggregation and linking of evolving structured, unstructured and streaming
data with lineage tracing
o Data Quality: Health data system 100 provides data quality services
to help organizations QA their data to ensure it will be clean and
standardized enough to deliver insights. The health data system 100
can also enforces data quality standards mandated by regulation,
ecosystem covenants, or the like.
Data Services o Patient Matching Service: Health data system 100 can match
multiple records from different sources (e.g., Hospital A and
Hospital B) to the same master person record.
o Data Integration Services: From a source system to health data
system 100 to avoid building many point to point integrations. Data
flows are configurable and "plug and play" based on ecosystem
member collaboration.
[0083] As discussed herein, the health data system 100 architecture can be developed in a
flexible manner to connect to a variety of data sources and services. For example, as illustrated in
Fig. 2, such components can include:
Big Data Can enable storage, The Big Data Pathway/Storage Module 313 can be a
Pathway/Sto processing, and analysis technology component that enables the ingestion,
rage Module of unstructured data management, and analysis of massive unstructured
313 data sets (e.g., Twitter social media data, physician
notes, etc.) in an efficient manner, using a variety of
best-in-class packages such as Apache Flume,
Cloudera (Apache) Hadoop, F P Vertica, and the like.
These tools allow for the quick and cost-effective
aggregation and searching through many terabytes
and petabytes of data, and will be involved whenever
big data analysis or analytics are desired.
Health Level Can enable The HL7 transaction set can specify a common
7 (HL7) interoperability with language to be used when exchanging data with
Transaction existing provider data healthcare providers. In various embodiments, it is
(not shown services the defacto standard for clinical data transmission in
in Fig. 2) the US, as featured in the HHS / CMS Meaningful
Use guidelines, and is gaining adoption worldwide.
The health data system 100 therefore supports HL7
transactions as a part of agent 205 and API 314
transmissions with healthcare providers or other
stakeholders that require provider data.
[0084] By leveraging the health data system 100 to enable a community-oriented approach to
delivering care and managing chronic disease, various healthcare projects, have the potential to
advance the shared mission of the national healthcare transformation effort and the Affordable
Care Act (ACA). Specifically, the health data system 100 can have a beneficial impact on each
of the ACA's "Triple Aims": providing more effective care and improve patient outcomes,
through democratization of evidence-based care to where patients are while ensuring quality of
care; providing care more efficiently and reduce the cost of care delivery, by triage and
allocation enabled by integrated care delivery network with access to patient data for advanced
analytics; and provide more consumer-centric care and improve patient experience, through uses
of mobile and social platform to personalize care experience.
[0085] The described embodiments are susceptible to various modifications and alternative
forms, and specific examples thereof have been shown by way of example in the drawings and
are herein described in detail. It should be understood, however, that the described embodiments
are not to be limited to the particular forms or methods disclosed, but to the contrary, the present
disclosure is to cover all modifications, equivalents, and alternatives.
[0086] Additional applications can include:
A. Improving chronic disease management by optimizing distribution of care across a
network of care providers, retail clinics, and the home
Chronic disease management is a high cost, low return problem that plagues the healthcare
industry. The problem spreads across multiple therapeutic areas such as diabetes, respiratory
illness and cardio-vascular diseases. On an average two thirds of total healthcare spend is
attributed to chronic disease management. Data driven chronic disease management has
already shown tremendous potential in reducing cost and improving quality of life and the
health data system 100 takes it to the next level through innovative interventions and
consolidation of data.
1. System 100 is an integral part of a chronic disease management program, providing
the core data management, ingestion, and application hosting services. System 100
provides delivery channel services that help caregivers/providers access the tools and
data they need to provide care seamless across facilities, while patients will have
similar access to tools and services they need to improve adherence to treatment
regimens. Some key capabilities of the system 100 that enable the new ecosystem
business model are listed below:
■ The data gatherer: System 100 provides the agent modules 205 that capture
data from multiple sources, such as EMR (Electronic Medical Records),
unstructured transactional and research data (through Natural Language
Processing), financial data, patient reported information, and data streams
from medical devices or fitness wearables. Data is captured from multiple
organizations/sources and tagged accordingly.
o The data agent modules 205 scheduled to run at pre-defined
intervals and/or can be executed on demand
o Each data agent module 205 provides a collection of queries
that run against data source system 201 and extract pre-defined
set of data values
o The data from data source 201 is transformed by the switch
3 11 using pre-defined algorithms and rule-sets and the
transformed data values are stored in the common information
model 312 either on persistent basis or in transitive mode
o For feedback purposes, the system 100 provides another set of
templates/attributes/lists that are populated based on the
outcome of the queries performed against the common info
model 312 by the switch 3 11
o System 100 includes multiple out of the box agent modules
205. These agent modules 205 have been built to extract data
from many 3rd party EMR systems such as EPIC, CERNER,
Centricity etc. and wearable devices such as BP and glucose
monitors. The agent modules 205 are unique with respect to
clinical and non-clinical attributes that are being extracted, as
well as the optimized queries written for extraction. In
addition, the agent modules 205 provide a unique plug-andplay
capability wherein the need for configuration is minimized
in case standard installations of transactional systems are being
used by the client.
o System 100 contains a Natural Language Processing (NLP)
engine that auto-tags unstructured data thereby reducing the
effort of document-tagging by 20-30%.
o The platform allows the merging of structured, unstructured,
and social data that is unique in context of healthcare industry.
Data consolidator: System 100 performs significant amount of harmonization
across the various ingested data sets, leveraging its unique transitive and
preservative data model, and stores the cumulative analytical core data set as
well as data received through feedback loop.
o Switch 3 11 maintains multiple code books that describe the
translation of one coding scheme (terminology) to another
o These code books are used to translate the inbound data's
coding scheme to the common coding scheme established for
the installation.
o For example one of the coding book explains the translation
rules for diagnosis coded in ICD9 to SNOMED coding scheme
o In case no one-to-one coding map is available for a particular
entry in a coding scheme, the switch module 3 11 leverages
user-defined rules to resolve the conflict.
o The transitive and preservative scheme of data storage is
controlled through a set of user-defined rules that distinguishes
between the data to be stored and data to be passed along, at
the time of execution
o For example there might be a rule that states that all data
received from a particular source system needs to be stored and
from another source system, needs to be pushed forward in a
cumulative state.
o System 100 contains a data model, the common information
model 312, that is patient-centric and includes multiple
domains such as vitals, chronic, episodic, genomic, claims,
social, behavioral etc. The common information model 312 is
not only unique from attributes and relationship perspective but
is also driven by user defined rules for storage perspective.
o The unique "transitive and preservative" concept deployed
within system 100, allows the data administrator to control the
flow of the data and its associated storage, through a set of
editable rules. System 100 provides the data administrator
control of what data should be stored persistently and what data
should be flowed through on a temporal basis.
■ Data provider: API 314 establishes an access layer for any application that
needs to access the data managed in common info model 312. System 100
includes SDK (Software Development Kit) that contains API 314, pre-defined
methods, and interface programs.
o API 314 can either be a collection of simple overloaded
methods or complex web services
o A method or a service usually maintains the following
structure:
Collection/Data set API Name (parameterList)
Where, Collection/Data set is the return value provided by
the API to the calling program/ system
API Name is the name that is used by the calling
program to invoke the functionality delivered through the
method/service
Parameter List is a number of parameters that are
passed by value or passed by reference to the being-called
method in order to shape the outcome/return result from the
method/service
o API definitions (signatures) are exposed through a pre-defined
mechanism or through a pre-defined directory service, for
example WSDL.
o APIs are called/invoked by the programs and systems that
interact with the system 100
o The calling program passes the values for the exposed
parameterList and the API return the response in form of either
a collection or a dataset, based on the passed values.
o System 100 contains variety of services/methods, categorized
under distinct API 314, that provide functionality to support the
needs of the calling program/system. Some examples include,
API 314 to return a consolidated medical record for a given
patient, API 314 to expose results of an analytical query
performed against a pre-determined cohort of patient
population, API 314 to pass a secured message between two
end users of the system 100 etc.
o The SDK (Software Development Kit) containing all the API
314 and other access mechanisms that are exposed by system
100 are unique to the system 100 wherein the signatures of the
API 314, the overloaded input attributes and output collections
have been designed and deployed specifically for system 314.
■ Security: All data in the system 100, whether at rest or in motion, is protected
through a sequence of security protocols in the security module 309. Security
module 309 provides a role based dynamic, multi-form authentication and
authorization. Data access is granted based on market rules 310 that establish
authorizations based on data source system, organization source (and
associated rules), and identities of the consuming user and application.
o Data at rest: All data stored in common information module
312 is encrypted using the latest encryption standards. The
data placed in temporary areas, such as staging, is also
encrypted. The encryption key is maintained at a distinct
location in a double-blinded vault. Each user is assigned a set
of roles that are maintained by the administrator of the
platform. These roles provide the users to access certain
screens, reports, modules, and code pieces. Even the users
with same access grants may not see same data because of
behind-the scene linkage between individual data items and
authorized user-group.
o Data in motion: There are two primary areas for data in
motion: 1) the data being retrieved from source systems, and 2)
data being exposed to program/ systems that need it through the
API 314. Under both scenarios, the data flows through a preestablished
secured tunnel. That secured tunnel is a
combination of multiple rules sitting on top of baseline SSL
protocol. Each layer of the platform (corresponding to ISO
model) has its own distinct rule-set that augment the core SSL
protocols.
o A 7-layer security and access control that is in line with the
ISO model for software deployment. 1) Each of the ISO layer
is independently handled through the security module 309, 2)
the process/algorithm used to handle each ISO layer is
proprietary to the system 100.
o The double-blinded vault to store encryption keys is a unique
concept adopted from clinical trial industry
■ Server 120 is PHI Certified: Server 120 is pre-configured to manage PHI
(Protected Health Information) datasets through the use of latest and strictest
available security protocols such as CS4096. Server 120 is HITRUST
certified.
• Server 120 is the only PHI carrying system built exclusively for
Academic Medical Centers (AMCs) that is both HITRUST certified
and deploys CS4096 encryption.
■ Analytical engine: System 100 provides a core analytical platform that
provides baseline analytical queries as well as propensity matched cohorts that
allow for focused analysis out of vast data sets contained in the common
information model 312 using the API 314.
o Data flows from the data sources 201 into the common
information model 312.
o Data is harmonized by the switch module 3 11.
o The harmonized data is cumulated and/or de-normalized based
on the analytical and/or transactional requirements for the data
The de-normalized/cumulated data is stored in common
information model 312
Patient records are classified into various cohorts based on
multiple factors such as age, gender, ethnicity, disease
conditions, comorbidities, care program enrollment, medication
types etc.
Baseline inferences are drawn across these cohorts and stored
as generic trends
The generic trends are compared to 3rd party
benchmarks/standards and variances are stored in the platform
When a query is executed using API 314, one or more patient
cohorts are identified as target population for the query,
depending upon the parameters passed by the query
The associated patterns/trends are exposed to the calling
program through API 314
The variance between the trends and real life data is identified
and quantified based on a pre-defined scale
In addition the end-users are provided with the capability to
over-ride the generated inferences
The adjustment factors (based on either original patterns and
real life data variances or end user over-rides) are fed back into
common info model 312 and act as margin/probability
adjustment factor for future iteration of similar query execution
using API 314
For cognitive queries, the trends are extrapolated into future
and intangible factors such as typical behavioral reaction of a
physician to a specific case, are superimposed on the baseline
query response to provide almost real-world predictions
The system 100 provides dynamic and self-learning analytics,
i.e., system 100 allows the end users to establish a reverse data
flow from feedback perspective which in turn adjusts the
boundary conditions, cut-off levels, percentage allocations,
probability distribution and other mathematical equations to
impact the next iteration of the execution of the analytical
queries. The self-learning and self-adjusting behavior of the
analytical responses based on transactional feedback gathered,
is a unique to the system 100
Patient attribution: System 100 precision triage API 314 takes a complete
health profile of an individual and matches him or her to the best fit facility
and health service based on health needs, provider specialties and capacity,
insurance coverage, access/convenience, and other factors that influence value
(quality, cost, experience).
o As data is ingested from the data sources 201 into the health
server 120, each record is compared against a pre-defined,
user-controllable set of attributes to link the record with variety
of cohorts
o The cohort definition is pre-defined and is based on the
ultimate functionality need of the system.
o For example for a chronic care management system, the user
may want to break down all the patients into three categories:
high-risk, medium-risk or borderline, and low-risk patients.
For each of these classifications, the administrator may have
defined a set of boundary conditions or rules. For example for
a diabetes management program, the rule set could be as
simple as the value of FIBAIC; less than 5 classified as lowrisk,
5-6 as borderline, and above 6 as high-risk.
o All the ingested data is pushed through these cohort definitions
and each record is tagged against multiple cohorts thereby
attributing the patients to various categories
o The patients, thus attributed into various cohorts, are further
used by the analytical engine as target population for end-user
queries.
o System 100 establishes relationships between patients and
various care delivery aspects. These care delivery aspects
could vary from incentive program enrollments, care
coordination, pre-section for trials etc. The linkage allows the
patients to be put under various cohorts thereby simplifying the
later query execution as well as improving performance
tremendously.
Consumer engagement: The consumer engagement API 314 implements a
patient portal with customized content based on an individual's health profile,
including geo-located and personalized incentives to encourage healthy
behavior, game-based tools that engage consumers to manage their health and
wellness, and direct connectivity to their healthcare providers.
o The patient portal is an integrated gateway to be used by the
patient to interact with system 100
o Individual patients access the portal through a secured
mechanism that deploys multi-form authentication that could
contain biometric authentication also, such as finger-print
and/or retinal scan
o Once in the portal, the patient is presented with a dashboard of
their cumulative health, including a proprietary healthScore
o The healthScore is created based on patient's medical condition
through the use of proprietary algorithms that leverage variety
of factors such as patient's demographic data, vitals,
concomitant meds being consumed, comorbid disease
conditions, prior history, family background, and multiple
other factors.
o The portal also offers multiple choices to the patients including
ability to interact with the caregivers, scheduling appointments,
refilling prescriptions, validating and adjusting their care plans,
learning through provided/recommended educational material
etc
o The patient portal is unique from the integration and
harmonization perspective, i.e., it is a single gateway that
allows the patients to see all their authorized data at one spot.
The data includes patient's medical data, lab results, radiology
images and reports, prescription data, their wearable device
data as well as wave forms (streaming) from devices such as
EKG. All these devices and system of records may use
differing terminologies. System 100 translates them into a
common language and terminology.
o Another unique feature of the platform is that it renders the
same information through same portal across multiple delivery
channels such as smart devices, laptops, phones etc.
Clinical Decision Support System: The system 100 provider care delivery
service includes a provider portal with customized applications, patient
clinical and pathology data, clinical resources like comparator data from
public resources. These capabilities will provide a longitudinal account of
patient data to improve cooperation across caregivers and coordination across
care facilities.
o The CDSS module of the platform can be rendered one of the
two ways: 1) as a standalone application, and 2) as an
integrated extension of existing transactional system
o Under the second scenario, the CDSS module is usually
invoked through the click of a button within the core
transactional application being used by the end user, such as an
EMR system for a caregiver.
o The patient Id for the patient under investigation, is passed to
the CDSS module
o The CDSS module retrieves and harmonizes the entire
available medical history for the patient Id passed patient Id.
o The CDSS module harmonizes the patient medical history
against the common terminology and standards being executed
by the transactional system and presents the longitudinal record
o In addition the CDSS module can be configured to provide
recommendations associated with the care plan for the
therapeutic pathway under consideration. These
recommendations are based on the patient's individual traits as
well as industry benchmarks and best practices
o From a CDSS perspective, the system 100 is driven by a set of
behind-the-scene rules that have been devised by clinicians and
are used to correlate a patient's medical condition to one or
more pre-defined therapeutic pathways in order to establish the
future care plan.
o The CDSS module leverages multiple industry benchmarks and
best practices in order to make recommendations for finetuning
future care plans
■ Outcome rate: 10-20% health outcomes improvement as measured by core
clinical parameter associated with the disease condition. For example 10-20%
improvement on HBA1C measure for diabetes or PFT score for asthma/COPD
■ Cost reduction: 15-25% cost reduction as measured against TCM (Total cost
of Management) for the disease condition. This cost reduction is measured
through reduction in high-cost activities (such as trips to ER) calculated across
multiple intervention cohorts
■ Compliance rate: 20-30% improvement in compliance rate as measured not
by the orders filled but orders consumed by patient.
■ Access improvement: 20-25% improvement as measured through a weighted
combination of various factors, such as reduction in time to schedule
appointments, reduction in days for the appointment, reduction in days to
access lab results etc.
Standardizing cancer treatment to reduce care variability for an integrated cancer care
delivery network
1. Over one trillion dollars in U.S. health care expenditures are considered wasteful,
driven by inefficiency in care delivery, suboptimal care, and inadequate access.
Cancer care is one of the most constrained due to insufficient access to expert
knowledge to the general oncology practitioner, lack of evidence-based tools and
practices to minimize treatment variations, and an inability to engage with the patient
to efficiently manage their disease. Current reimbursement models do not incentivize
investments in long term wellness, instead focusing on treatment volume. These
factors create gaps in treatment quality and increase costs. A new health system is
needed to deliver enduring changes enabled by innovative data-driven solutions that
demonstrate improved outcomes and cost savings at-scale. The MD Anderson Cancer
Center (MDACC) Network Democratization project is one such initiative, committed
to accelerating clinical care for cancer patients by improving outcomes through
evidence-based and experience-informed clinical decisions.
2 . The network democratization initiative focuses on the following broad level activities
■ Creating of an integrated provider partner ecosystem that consists of Providers
(to manage patients), Risk Bearing Entities (that pay for the care), and
Technology partners (to enable allocation of the right care, facilitate care
continuity, and improve care quality)
■ Securely ingesting data from multiple sources (such as EMR systems) for
cancer patients. Standardizing clinical data into a common format and
terminology to create a curated disease summary, identifying treatment
recommendations for demographically similar cohorts based on expert
knowledge, enabling expert virtual consults for diagnosis and staging
confirmation, and providing care pathway advisories to manage the disease
■ Creating avenues for targeted interventions through education and cancer
screening services
■ Shifting preventative screening and early diagnosis to the community
providers (60%), personalized care supported by expert advisory and live
expert consultation (30%), and facilitating rapid referrals if the expertise is not
available at the local network partner (10%)
■ Making available consistent and consolidated information to all involved
stakeholders (physician, support personnel, patient) through web and mobile
solutions
■ Leveraging a longitudinal view of the patient's cancer and medical history
data to fine-tune treatment recommendations
■ Comparing cost and outcomes for intervention programs against nationally
and/or internationally established benchmarks, to further tune interventions
for the right outcomes
System 100 is an integral part of network democratization, providing core data
management, data ingestion/ standardization, and analytics hosting.
■ External Data Connectors and Services: The agent modules 205 capture
data from multiple sources, such as EMR (Electronic Medical Records)
systems resident at a network provider partner, unstructured clinical data
(used to ingest secondary information about the patient), and patient selfreported
information (such as a cancer survey completed during a hospital
onboarding event)
o The agent modules 205 can be scheduled to run at pre-defined
intervals and/or can be executed on demand.
o Each agent module 205 includes queries that run against the
source 201 to extract pre-defined set of data values.
o These data attributes can be controlled through templates that
come out of the box with system 100. These templates can be
modified at installation time or during execution by an
administrator of system 100.
o The extracted data is transformed using pre-defined algorithms
and rule- sets and the transformed data values are stored in 312
o System 100 includes multiple out of the box agent modules 205
that are proprietary to the platform. These agent modules 205
have been built to extract data from both custom and 3 party
EMR systems such as EPIC, CERNER, etc. and wearable
devices such as BP and glucose monitors. The agent modules
205 are unique with respect to clinical attributes that are being
extracted, the proprietary two-step process being used for
extraction and transformation, as well as the optimized queries
written for extraction. In addition, the agent modules 205
provide a unique plug-and-play capability wherein the need for
configuration is minimized in case standard installations of
transactional systems are being used by the client
o The system 100 allows the merging of structured and
unstructured data that is unique in context of care delivery
■ Data Management: System 100 standardizes data across the various data sets
to a variety of clinical data exchange standards, leveraging its unique patientcentric
common information model, and persists primary clinical data, as well
as derived secondary information (e.g., prognostic scores) collected from
clinical notes (e.g., secondary clinical diagnoses)
o System 100 maintains multiple coding dictionaries that
describe the translation of one coding scheme (terminology) to
another.
o These coding dictionaries are used to translate the terms in the
data from data source 201 into a set of standard terms used in
system 100.
o For example one of the coding dictionaries contains translation
rules for clinical diagnoses coded in ICD9 to SNOMED coding
scheme. Another example translates lab data into the LOINC
coding scheme
o In case no one-to-one coding map is available for a particular
entry in a coding scheme, system 100 leverages user-defined
rules to resolve the conflict. System 100 persists both the as-is
and translated values
o The transitive and preservative scheme of data storage is
controlled through a set of user-defined rules in 310 that
distinguish between the data to be persisted in 312 and data to
be obtained dynamically from 201 at the time of a request for
this data by 130
o For example there might be a rule that states that all data
received from a particular source system needs to be persisted
in 312 and from another source system, needs to be obtained
from 201 on demand.
o System 100 contains a data model, common information model
312 that is patient-centric and includes multiple domains such
as demographics, home medications, allergies, vitals,
observations, labs, and clinical document relevant to cancer
care. 312 is not only unique from attributes and relationship
perspective but is also driven by user defined rules for storage
perspective.
o The unique "transitive and preservative" concept in system
100, allows data source owner and / or administrator of system
100 to control the amount of data from data source 201 that is
persisted in 312 by an administrator of system 100.
■ Security: System 100 uses a role based dynamic, multi-form authentication
and authorization protocol 310. Data access is granted based on rules that
establish authorization based on data source, and identities of a consuming
user (e.g., a Lung physician at a Community hospital) and consuming
application (e.g., Oncology Expert Advisor analytic tool). System 100 is preconfigured
to manage PHI (Protected Health Information) datasets through the
use of latest and strictest available security protocols such as CS4096. In
addition, the system 100 is HITRUST certified
■ Access, Discovery, and Analytics: System 100 contains over 200 API 314
purpose-built for network democratization, to establish an information
delivery and access layer for consuming analytics tools.
o API 314 can either be a collection of simple overloaded
methods or complex web services
o A method or a service usually maintains the following
structure:
Collection/Data set API Name (parameterList)
Where, Collection/Data set is the return value provided by
the API to the calling program/ system
API Name is the name that is used by the calling
program to invoke the functionality delivered through the
method/service
Parameter List is a number of parameters that are
passed by value or passed by reference to the being-called
method in order to shape the outcome/return result from the
method/service
o API definitions (signatures) are exposed through a pre-defined
mechanism or through a pre-defined directory service, for
example WSDL.
o API 314 are called/invoked by the user devices / services 130
that interact with system 100.
o The calling program passes the values for the exposed
parameterList and the API return the response in form of either
a collection or a dataset, based on the passed values.
o System 100 contains variety of services/methods, categorized
under distinct API 314, that provide functionality to support the
needs of the calling program/system. Some examples include,
API to return a consolidated medical record for a given patient,
API to expose results of an analytical query performed against
a pre-determined cohort of cancer patients, or an API to pass a
secured message between two network partners using system
100
■ Key Outcomes: Reduced variability in cancer treatment, improved diagnosis
and staging accuracy, and better adherence to care pathways leading to
reduced unnecessary ER/hospital visits and empowered patients with
education, screening, and adherence tools
■ Cost Reduction: Lung cancer creates over 220K new patients in the US each
year, with an estimated $8 - $10B spend on treatment of non-small cell lung
cancers (NSCLC) annually (treatments are 50% more expensive for patients
over 65 years of age). The outcomes indicated above represent a 16% savings
opportunity in the overall cost of lung cancer treatment, translating to a $180 -
$240M reduction in cost of care for NSCLC (or 85% of new lung cancer cases
diagnosed each year)
■ Breast cancer creates over 23OK new patients in the US each year, with an
estimated $16 - $17B spend on treatment (treatment at stage 3 and 4 of breast
cancer can cost up to five times more than stages 1 and 2). The outcomes
indicated above represent an 8 - 12% savings opportunity in the overall cost
of breast cancer treatment, translating to a $192 - $204M reduction in cost of
care assuming a 5 - 10% patient engagement rate (of the overall population)
Increasing therapy adherence to improve chronic disease outcomes for pharma
companies
1. Medication non-adherence represents a $300B challenge to the pharma industry -
leading to 125K deaths each year (4th leading cause of death, 1st for accidental
deaths), causing 10 - 20% of hospital or nursing home admissions, and $188B in
biopharma revenue loss. Today's adherence programs solve for this challenge
through limited patient communication via traditional channels (mail/email/call), are
not based on real-time data about dosing behavior, limit health care professional's
(HCP) ability to link adherence to symptom improvements, and are unable to
continuously improve efficacy of messages to patients. In response, Pharmas are
launching digital and analytics enabled solutions to personalize interventions and
demonstrate improved outcomes through increased patient engagement.
System 100 plays a unique role in helping pharma companies play a role in actively
monitoring patient symptoms, stay on recommended therapy regimens, and reinforce
healthy behaviors (e.g., getting a refill, seeing a clinician). System 100 facilitates
secure continuous ingestion, processing, and monitoring of informed patient consent,
combines sensitive information (PHI) related to dosage, combining it with other
patient information such as insurance claims, clinical records, and self-reported
symptoms. This data is used by analytic algorithms to draw inferences such as
propensity scores, adherence patterns, and outcome metrics to stratify patients by risk
and recommend interventions; patient and provider tools then communicate with the
patient or HCP with targeted interventional and/or educational content to improve
adherence behaviors and effect outcomes.
■ Creating an integrated partner ecosystem comprising Health Care
Professionals (HCPs that treat patients), Risk Bearing Entities (insurers that
facilitate Medical and/or Pharmacy claim processing), Patients, and
Technology partners (device partners to signal adherence data, partners to
enable personalized mobile communication, analytics partners to generate
individual or cohort-level insights)
■ Securely ingesting and harmonizing data from multiple sources into a
common format to stratify at-risk individuals, understand what interventions
work for whom and via what channel, and customize messages for therapyrelated
engagement
■ Facilitating targeted interventions (messages, in-app alerts, educational
content, dashboards) for the Patient and/or HCP to supplement existing
reminders from specialty pharmacies
■ Conducting post market studies to demonstrate impacts of improved
adherence
System 100 is an integral part of adherence solutions, providing secure data
collection, standardization, and analytics hosting.
Analytical engine: System 100 provides a core analytical platform,
implemented as a of services provided by the API 314, that provides baseline
analytical queries as well as propensity matched cohorts that allow for focused
analysis out of heterogeneous data sets
o Data flows from source systems into system 100.
o Through the data gatherer and data consolidator roles, the data
is ingested and harmonized
o The harmonized data is cumulated and/or de-normalized based
on the analytical and/or transactional requirements for the data
o The de-normalized/cumulated data can be stored in specific
data marts for targeted analysis
o Patient records are classified into various cohorts based on
multiple factors such as age, gender, ethnicity, disease
conditions, comorbidities, care program enrollment, medication
types etc.
o Baseline inferences can be drawn across these cohorts and
stored as generic trends that can be used to customize and
recommend interventions
o The generic trends are compared to 3 party
benchmarks/standards and variances are stored in the platform
o When a query is executed, one or more patient cohorts are
identified as target population for the query, depending upon
the parameters passed by the query
o The associated patterns/trends are exposed to the calling
program through API 314
o The variance between the trends and real life data is identified
and quantified based on a pre-defined scale
o In addition the end-users are provided with the capability to
over-ride the generated inferences
o The adjustment factors (based on either original patterns and
real life data variances or end user over-rides) are fed back into
system 100 platform and act as margin/probability adjustment
factor for future iteration of similar query execution
o For cognitive queries, the trends are extrapolated into future
and intangible factors such as typical behavioral reaction of a
physician to a specific case, are superimposed on the baseline
query response to provide predictions
o System 100 offer dynamic and self-learning analytics, i.e.,
system 100 allows the end users to establish a reverse data flow
from feedback perspective which in turn adjusts the boundary
conditions, cut-off levels, percentage allocations, probability
distribution and other mathematical equations within the
analytical engine in order to impact the next iteration of the
execution of the analytical queries. The self-learning and selfadjusting
behavior of the analytical responses based on
transactional feedback gathered, is a unique preposition of the
system 100.
■ Information delivery gateway: The system 100 consumer engagement
service includes a patient portal with customized content based on an
individual's health profile, including geo-located and personalized incentives
to encourage healthy behavior, game-based tools that engage consumers to
manage their health and wellness, and direct connectivity to their healthcare
professional (HCP)
o The system 100 patient portal is unique from the integration
and harmonization perspective, i.e., it is a single gateway that
allows the patients to see all their authorized data at one spot.
The data includes patient's medical data, lab results, radiology
images and reports, prescription data, their wearable device
data as well as wave forms (streaming) from devices such as
EKG. All these devices and system of records may use
differing terminologies.
o Another unique feature of the platform is that it renders the
same information through same portal across multiple delivery
channels such as smart devices, laptops, phones etc.
■ Facility for end-to-end sensitive patient data management: System 100
uses a role based dynamic, multi-form authentication and authorization
protocol 309. Data access is granted based on rules that establish
authorization based on data source, and identities of a user {e.g., Patient,
Pharma, HCP user), system {e.g., EMR system providing patient clinical
data). System 100 is pre-configured to manage PHI (Protected Health
Information) datasets through the use of latest and strictest available security
protocols such as CS4096. In addition, the system 100 is HITRUST certified
Improving financial and spend performance using advanced analytics and multi-data
collection leading to activity based costing
1. Large scale institutes (Academic as well as pure play commercial) with multiple
facilities spread across geographically, usually are forced to manage a diverse and
complicated supply chain. The "spend" is usually highly distributed with a lack of
centralized coordination, leading to significant redundancy as well as wastage. The
critical challenges faced under such situations range from under-utilization of
inventories at one facility whereas same inventory is in short supply at another
facility, identification of best vendors for local quick-lead supplies as well as
centralized long-lead needs, Economic Order Quantity and reorder level identification
etc. These supply chains are also encumbered by the fact that usually they are a
network of multiple locally operated ordering and procurement systems, thereby
highly unlikely to track "spend" across an individual patient's life cycle.
There is an incremental trend that the spend analytics is moving towards and that is to
establish a solid platform to calculate and manage 'activity based costing' in order to
support the latest healthcare paradigm of bundled payments. It is critical for any large
institution to understand the details of cost broken down by each involved activity
under a large scale procedure (such as hip replacement) for them to be able to
logically accept an offered bundled payment for the procedure by the RBEs.
Any well-tuned centralized spend analysis system focuses on following broad level
activities
■ Creation of an integrated ecosystem that consists of suppliers (local as well as
centralized), inter-linked facilities that service the patients and a centralized
governance authority to manage the flow of data across all stakeholders and to
manage the global rules that govern procurement.
■ Ingestion of data from multiple local ordering and procurement systems
■ Ingestion of data from local clinical and operational systems such as EMRs,
scheduling systems, pharmacy systems, financial systems, and capacity
management systems.
■ Ingestion of data from third party sources that provide benchmarks and best
practices across the country.
■ Harmonization of "spend" and clinical data to generate a common script for
the centralized authority.
■ Spend data categorization across multiple factors such as type of
material/services, geographical demand and availability, clinical necessity and
urgency, capital or non-capital needs etc.
■ Tying the spend data to clinical and/or therapeutic pathways in order to
generate activity based cost for each sub-activity of a large procedure. This
information is critical for the newer payment models coming up in the health
industry, such as payment for outcomes rather than payments for service.
■ Making available consistent and consolidated information to all involved
stakeholders (local facilities, suppliers, centralized authority etc) through
various delivery channels
■ Establishing spend patterns, trends, lead times, replenishments etc based on
harmonized spend data
■ Comparison of spend trends against 3 party benchmarks to identify the
variances and associated mitigation approaches
■ Ability to establish rules (user controlled) to cover for cost versus clinical
urgency requirements
■ Ability to establish rules to manage and support local variances
■ Ability to receive feedback and incorporate it in order to fine-tune future
transactions
■ PHI Certified System 100: System 100 is pre-configured to manage PHI
(Protected Health Information) datasets through the use of latest and strictest
available security protocols such as CS4096. In addition the system 100 is
HITRUST certified.
• Uniqueness: System 100 is the only PHI carrying platform built
exclusively for Academic Medical Centers (AMCs) that is both
HITRUST certified and deploys CS4096 encryption.
■ The system 100 patient portal is unique from the integration and
harmonization perspective, i.e., it is a single gateway that allows the patients
to see all their authorized data at one spot. The data includes patient's medical
data, lab results, radiology images and reports, prescription data, their
wearable device data as well as wave forms (streaming) from devices such as
EKG. All these devices and system of records may use differing
terminologies. System 100 translates them into a common language and
terminology.
o Another unique feature of the platform is that it renders the
same information through same portal across multiple delivery
channels such as smart devices, laptops, phones etc.
■ Activity based costing engine: System 100 ties financial data with clinical
data to establish spend patterns against each sub-activity of a large therapeutic
procedure. For example system 100 can break down spend against pre-op,
day-of-op, and post-op care activities for a typical hip replacement procedure
that is neither chronic nor acute care but instead spreads across a pre-defined
time-frame.
o System 100 contains activity level breakdown of multiple
therapeutic pathways for a variety of disease conditions. These
pathways are neither chronic not episodic. Instead they focus
on a period of time wherein patient was treated for a particular
disease condition or abnormality.
o For example, there is a detailed breakdown of all pre-op,
operational, and post-op activities associated with hip and knee
replacement process.
o Post financial-data-ingestion, the platform tags the
transactional data against these activities that cumulate to
formulate the complete process
o This categorization of financial transactions is cumulated
across multiple patients and multiple facilities
o Analytics is performed against these cumulative numbers to
establish factors such as average cost per activity for a
particular therapeutic pathway, best facility to perform a
certain activity based on cost as well as outcome rates etc.
o These anal
■ Contract support engine: System 100 leverages the trends established
through activity based costing to provide analytical support for providers as
well as RBEs in order to establish contract terms for newer payment models
such as value based payments and/or bundled payments.
o As an activity based costing engine, system 100 breaks down
certain therapeutic pathways into sub activities.
o For example there is a clear list of pre-op, day-of-op, and postop
activities that have been defined for a hip replacement
procedure. These activities could span a considerable amount
of time but are not considered chronic conditions.
o It is imperative to understand the cost associated with each of
these sub-activities in order to establish contracts under new
payment models for the healthcare industry. Models such as fee
for value rather than fee for service or bundled payment
contracts.
o System 100 leverages baseline benchmarks created under
activity-based-costing module, for each sub-activity,
considering variety of factors such as patient-mix, geographical
and environmental correlations, and health outcomes, and helps
the financial administrators in figuring out what is the optimum
contract terms that they can establish with risk bearing entities,
for a given therapeutic pathway.
■ Capacity Utilization: 10-15% improvement on capacity utilization based on
centralized procurement based on system 100 suggested EOQ and
replenishment levels. This is also impacted by the geographical distribution of
the facilities as well as extent of centralized procurement.
■ Contract Efficiency: 10-12% higher accuracy in terms of matching actual
spend against the offered reimbursement by RBEs. These matrices are also
impacted by the volume of such bundled payments as well as clearly
identified sub-activities within a therapeutic pathway.
■ Patient-mix based capacity adjustment: 15-20% improvement in capacity
utilization as measured against the patient-mix at a facility. System 100
provides recommendations, based on patient-mix as well as volumes,
regarding the optimum capacity for each facility, leading to the
aforementioned reduction in wastage and/or over-stocking
■ Administrative cost efficiency: 12- 14% cost efficiency (reduction in overall
spend system cost against system 100 cost) is a typical ROI that can be
achieved through system 100
ating translational cancer research at a comprehensive cancer care center
Cancer is the second leading cause of death in the USA with almost 600,000
Americans succumbing annually and generating $175 billion in healthcare system
cost. As a result of decades of investment in cancer research, the life expectancy for
many of the 1.6 million Americans diagnosed with cancer every year is steadily
increasing. However, much remains to be done as there are still relatively few true
cancer cures, and some cancer types have seen little improvement in mortality over
the last few decades. Cancer is now recognized as essentially a disease of the
genome. DNA damage leads to uncontrolled cell division and often rapid cellular
evolution which enables cancerous populations to develop resistance to therapy.
There is a big push to translate the latest breakthroughs in the genomic sciences to
cancer care delivery to create personalized medicine approaches tailored to each
patient's particular cancer genome. This typically involves integrating large scale
genomic and clinical data on the order of many gigabytes per patients and petabytes
across even relatively small cancer cohorts.
Translational cancer research typically involves:
■ Acquiring vast amounts of mic data on each patient including genomic,
epigenomic, RNA sequences, and proteomics data from both the germline and
the cancer as it evolves over time.
■ This patient-level mics data must be combined with other laboratory assays
and rich clinical data from the patient's medical record including diagnoses;
surgical, radiation, and chemotherapy; responses to treatment, and, ultimately,
outcomes.
■ Once these data are integrated for each patient, cancer researchers query the
data looking for gene mutations and other biomarkers that predict prognosis
and response to therapy.
■ These incites are useful for generation hypotheses about cancer biology, new
treatment modalities, and cancer population health as well as to design clinical
trials.
■ In some cases, this integrated data can also inform actually clinical therapeutic
chooses for individual patients, i.e. personalized medicine
3 . System 100 is an integral part of a translation cancer research program, providing the
core data management, ingestion, and application hosting services. System 100 is
also provides delivery channel services that help cancer researchers access the tools
and data they need to drive their research in the lab, at the bedside, and across patient
populations. Some key system 100 capabilities that enable the new ecosystem
business model are listed below:
Increasing therapy adherence to improve chronic disease outcomes for pharma
companies
The Information Interchange helps leading pharma companies play a unique role in using
data and analytics to provide tools directly to patients and their health care provider teams
that help them monitor symptoms, stay on the recommended therapy regimen, and reinforce
healthy behaviors (e.g., getting a refill, seeing a clinician). The Interchange facilitates secure
continuous ingestion, processing, and monitoring of informed patient consents, and combines
sensitive information (PHI) related to dosage from wearables (e.g., Fitbit, Apple Watch),
drug injectors, insurance claims, clinical profiles, and self-reported health and activity from
patient mobile applications.
The sensitive information is then used by analytics algorithms for inferences (e.g., propensity
scores, adherence patterns, outcomes metrics) to stratify patients by risk, and to recommend
interventions based on predicted behaviors; patient and provider-specific tools then inform
the patient and their clinician with targeted interventional and informational content.
Improving Product Quality and Safety of Imported Drugs
The system 100 can serve as an information interchange that will enable the pharmaceutical
industry to satisfy the federal expectations set forth in Title VII of the Food and Drug
Administration Safety and Innovation Act (FDASIA) (Pub. L. 112-144). Title VII provides the
FDA new authorities to help ensure the safety, effectiveness and quality of drugs in the United
States. For example, the information interchange could help with the following subset of the 18
sections of Title VII:
1) Sec 705: Risk-based Inspection Frequency. Off shore drug manufacturers will need to
submit data periodically to the FDA in order to determine a risk-based schedule of
inspections of off-shore facilities with the finite number of inspectors available. Thus, the
interchange will enable manufacturers to integrate product, facility, raw material, and
operator data readily for evaluation by quality system analytics. This ability to produce
product quality data periodically not only enables manufacturers to comply with this
section of Title VII, but also enables review of their quality systems and to take proactive
action prior to an inspection.
2) Sec 710: Exchange of Information. FDA may enter into reciprocal agreements to furnish
certain trade secret information to foreign governments that have the authority and
demonstrated ability to protect trade secret information. The HITRUST certified
information interchange is uniquely positioned to support handling of trade secrets and
other proprietary aspects of product quality or compound composition data.
3) Sec 7 11 : Enhancing the Safety and Quality of the Drug Supply. The information
interchange can improve the oversight and control that manufacturers demonstrate over
the manufacture of quality drugs, including raw materials and finished products. See
diagram below of quality systems data that interchange may monitor and integrate to
support this reporting requirement.
CLAIMS
What is claimed is:
1. A health data system for delivering patient-centric and value-based care,
comprising:
a health data server;
one or more health data sources in communication with the health data server over a
secured network, wherein said health data sources each have a set of polling permissions;
one or more agent modules of the health data server that poll health data from the data
sources at a designated frequency based on a set of identifiers and the set of polling permissions;
a first switch module for providing the polled health data into a common information
model, the common information model being defined by at least one patient record, each patient
record having one or more attributes; and
one or more interface modules for gaining access to the common information model
based on a set of access permissions.
2 . The health data system of claim 1, wherein the common information model
includes a distributed database and wherein the one or more attributes optionally define at least
one of clinical health data, laboratory data, remote monitoring data, biometrics, wearables, social
media data, self-reported data, mobile application data, and device instrumentation.
3. The health data system of claim 1 or claim 2, wherein said first switch module
augments the common information model with a new attribute when the polled health data does
not map into the one or more attributes.
4 . The health data system of any one of claims 1-3, wherein said first switch module
at least one of filters the polled health data prior to providing the polled health data into the
common information model based on storage permissions, the storage permissions optionally
being provided by at least one of the health data server and the one or more health data sources,
matches patient records against each other, and controls network connections over the secured
network.
5 . The health data system of any one of claims 1-4, wherein the designated
frequency is set by at least one of the health data server and the one or more health data sources.
6 . The health data system of any one of claims 1-5, further comprising a second
switch module for providing the polled health data into the common information model, wherein
said first switch module communicates with said second switch module for receiving the polled
health data.
7 . A method for delivering patient-centric and value-based care, the method
comprising:
polling one or more health data sources for health data via one or more agent modules of
a health data server, wherein each of the one or more health data sources have a set of polling
permissions;
populating a common information model with the polled health data via a first switch
module, the common information model being defined by at least one patient record, each patient
record having one or more attributes; and
providing access to the common information model via one or more interface modules
based on a set of access permissions.
8 . The method of claim 7, wherein said populating the common information model
comprises populating a distributed database optionally with the one or more attributes selected
from at least one of clinical health data, laboratory data, remote monitoring data, biometrics,
wearables, social media data, self-reported data, mobile application data, and device
instrumentation.
9 . The method of claim 7 or claim 8, further comprising augmenting the common
information model with a new attribute when the polled health data does not map into the one or
more attributes via the first switch module.
10. The method of any one of claims 7-9, further comprising filtering the polled
health data based on a set of storage permissions prior to said populating.
11. The method of any one of claims 7-10, wherein the set of storage permissions are
provided by at least one of the health data server and the one or more health data sources.
12. The method of any one of claims 7-1 1, wherein said polling occurs at a
designated frequency set by at least one of the health data server and the one or more health data
sources.
13. The method of any one of claims 7-12, further comprising matching patient
records via the first switch module.
14. The method of any one of claims 7-13, further comprising polling a second switch
module via the first switch module for the polled health data.
15. The method of any one of claims 7-14, wherein said polling is limited by the set
of polling permissions.