Clinical Software Architectures · well as ordering of drugs. POE systems support formulation of...
Transcript of Clinical Software Architectures · well as ordering of drugs. POE systems support formulation of...
Clinical Software Architectures
Benjamin Bergner
Data Management for Digital HealthWinter term 2018/19
Agenda
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
2
DataSources
DataFormats
BusinessProcesses
Processing and
Analysis
Oncology Nephrology Intensive Care
Additional Topics
Dat
a M
anag
emen
t&
Fou
ndat
ions
Rea
l-w
orld
U
se C
ases
BiologyRecap
Agenda
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
3
DataSources
DataFormats
BusinessProcesses
Processing and
Analysis
Oncology Nephrology Intensive Care
Additional Topics
Dat
a M
anag
emen
t&
Fou
ndat
ions
Rea
l-w
orld
U
se C
ases
BiologyRecap
■ Only 60% of German clinics use EMR for internal access1
■ Germany occupies 10th place among 20 EU countries2
■ Leaders: Denmark, Sweden and Estonia
■ In Denmark: 99% of discharge letters are digital3
Electronic Medical Records (EMR) in Germany
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
4[1] https://www.presseportal.de/pm/8664/2723540[2] http://www.stiftung-muench.org/studie-zur-elektronischen-patientenakte-im-ausland-klare-vorgaben-des-gesetzgebers-sind-voraussetzung-fuer-erfolgreiche-implementierung/[3] http://www.sum.dk/~/media/Filer%20-%20Publikationer_i_pdf/2012/Sundheds-IT/Sundheds_IT_juni_web.ashx[4] https://e-estonia.com/solutions/healthcare/e-health-record/
Stiftung Münch, study of electronic
patient records in Europe (2015)4
■ Measures the adoption and utilization of EMR functions required to achieve a paperless environment that harness technology to support optimized patient care.
■ Shows progress against other healthcare organizations
EMR Adoption Model (EMRAM)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
5
http://www.himss.eu/healthcare-providers/emram
EMR Adoption Model (EMRAM)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
6http://www.himss.eu/healthcare-providers/emram
EMR
AM
sta
ge 6
/ 7
aw
arde
es
Economic Benefits of Healthcare Digitalization in Germany
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
7
https://www.mckinsey.de/publikationen/2018-09-27-digitalisierung-im-gesundheitswesen
■ Electronic Medical Record (EMR):
□ Digitized versions of paperwork in a clinician’s office including medical history, diagnoses, medications, immunization dates, allergies
□ Transferring data out of the practice from EMRs is not convenient; patient records have to be printed or mailed for consultations.
■ Electronic Health Record (EHR):
□ Focus is on a patient’s total health, not just the standard clinical data
□ Broader view of care including overall past medical history, EMR data, lab data, imaging reports.
□ Can contain other relevant information such as insurance information, demographics, even data imported from personal wellness devices.
□ Facilitates data sharing outside the practice with other health care providers such as laboratories and specialists.
EMR vs EHR
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
8
Processed Data in Hospitals: Patient Care
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
9
Entity type Description
Patient is a person being subject of care; information about a patient includes the patient identification number (PIN)
Case mostly comprises a patient’s stay in a hospital from patient admission to patient discharge or several ambulatory treatments related to one disease; information about a case includes the case identification number (CIN)
Order is a request for a diagnostic, therapeutic or drug service, e.g. a laboratory order or a radiological order
Diagnosis is the identified cause or nature of a disease or medical condition
Processed Data in Hospitals: Resources
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
10
Entity type Description
Appointment determines what persons have to be at a certain place at a given time. Examples are appointment for patient admission, examination or surgery.
Bed must be managed according to its occupation.
Health care professional
a health care professional who treats according to his or her specialization (e. g. nephrology or pediatrics) patients with certain diagnoses. Health care professionals are for example physicians and nurses.
Drug is a substance administered to a patient for treatment, diagnosis orprevention
RxNorm - Contains all Medications Available on US Market.
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
11
Processed Data in Hospitals: Administration
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
12
Entity type Description
Patient record archive
describes how and where the electronic or paper-based patient record can be found.
Classification of diagnoses
e.g. the International Classification of Diseases (ICD).
Cost unit information about a person or an institution responsible for bearing the costs
Processed Data in Hospitals: Management
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
13
Entity type Description
Business strategy defines the hospital’s long-term strategic goals
Strategicinformationmanagement plan
a strategic plan, which gives directives for the construction and development of a hospitalinformation system
Project a unique undertaking that is characterized by management objectives, restrictions with regard to available time and resources, and a specific project organization (DIN 69901)
Quality report openly published report about a hospital’s performance
Main Hospital Enterprise Functions
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
14
■ Patient admission (short: admission) aims at recording and distributing the patient demographics and insurance data as well as medical and nursing data of the patient history. Each patient must have a unique identifier and be assigned to a unique case.
-------------
■ Appointment scheduling, also in case of emergencies
■ Patient identification and checking for recurrent, using health insurance card
■ Administrative admission, creates a case summarizing hospital stay
■ Medical admission, done by physician, comprises patient history, admission diagnosis is stated as a result
■ Nursing admission, information like communication ability, nutrition, mobility
■ Visitor and information services, where is patient located?
Patient Admission
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
15
■ Decision must be made upon diagnostic and therapeutic procedures
■ Must be able to consult internal/external experts, e.g. with tumor boards in cancer treatment
Decision Making, Planning, Organization of Patient Treatment
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
16
Example for Planning: Molecular Tumor Board Software
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
17
■ Preparatory work like researching on medications for mutations from literature and (curated) databases
Example for Planning: Molecular Tumor Board Software
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
18
■ Presentation of patient research during the meeting
Example for Planning: Molecular Tumor Board Software
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
19
■ Documentation of decided upon medication that can be given back to the treating physician
Example for Planning: Molecular Tumor Board Software
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
20
■ Follow-up of treated patients for cancer registry
■ Decision must be made upon diagnostic and therapeutic procedures
■ Must be able to consult internal/external experts, e.g. with tumor boards
■ Medication prescription may be supported by providing knowledge about adverse drug events
Decision Making, Planning, Organization of Patient Treatment
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
21
Canada Vigilance Adverse Reaction Online Database
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
22
■ Decision must be made upon diagnostic and therapeutic procedures
■ Must be able to consult internal/external experts, e.g. with tumor boards
■ Medication prescription may be supported by providing knowledge about adverse drug events
■ Diagnostic/therapeutic procedures should be explained to the patient, documentation of the patient‘s informed consent and the decision making
Decision Making, Planning, Organization of Patient Treatment
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
23
■ Diagnostic and therapeutic procedures must often be ordered at specialized service units (e.g. laboratory, radiology, or pathology). These units execute the ordered procedures and communicate the findings or results back to the ordering department
■ Preparation of an Order: specimen collection and assignment to patient
■ Appointment Scheduling, e.g. time slot for image modality, patient transport
Order Entry
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
24
Paper Order Entry
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
25Paper-based order entrysystem for laboratory testing
■ Hospital must plan to offer adequate tools/resources (staff, room, equipment) for the execution of the necessary procedures.
■ All clinically relevant patient data (vital signs, orders, results, decisions) must be recorded and made available among involved persons, but also for legal justification.
■ Findings and reports must be transmitted to the ordering unit as fast as necessary
■ Data should be recorded in a structured way enabling statistics, computerized decision support and data retrieval
■ Data must always be linked to PIN and CIN
Execution of Diagnostic, Therapeutic and Nursing Procedures
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
26
■ The hospital must be able to document and code all diagnoses stated
■ Are the basis for the hospital's billing!
■ Coding in a standardized way ICD-10
Coding of Diagnoses and Procedures
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
27
International Classificationof Diseases (ICD)
■ Patient is discharged after treatment is terminated. Sometimes, patient is referred to other institutions
■ Administrative Discharge: Final billing happens, based on Diagnosis Related Groups (DRG): Criteria like diagnosis, procedures, patient age used for calculation
Patient Discharge
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
28Billing in Germany
■ Patient is discharged after treatment is terminated. Sometimes, patient is referred to other institutions
■ Administrative Discharge: Final billing happens, based on Diagnosis Related Groups (DRG): Criteria like diagnosis, procedures, patient age used for calculation
■ Medical Discharge: Attending physician writes a medical report including:
□ The relevant diagnosis
□ Important findings
□ Therapeutic procedures
□ The current patient state
□ Recommandations for further treatment
■ Nursing Discharge: Report is written including information like activity level, diet, wound care, etc.
Patient Discharge
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
29
■ Hospital Information Systems:
□ Enables administrative functions around patient billing
■ Clinical Information Systems:
□ Enables patient management (e.g. documentation, scheduling)
■ Decision Support Systems:
□ e.g. regarding diagnoses, treatment decisions
Different Terms for Software Components
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
30
■ Especially used for patient admission, discharge and billing
■ It must provide correct, complete and up-to-date administrative patient data for all other application components
■ In addition, all other application components must be able to transmit relevant administrative patient data (e.g., diagnoses) to the patient administration system
Application Components: Patient Administration System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
31
Application Components: Patient Administration System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
32
Screenshot of a patient administration system showing the patient list of a department of neurosurgery in the background and a window for assigning an appointment to a patient in the front
■ Supports specific documentation tasks with modules for different medical fields
■ Provides functions like speech-to-text, reuse of already documented data
■ Coding of diagnoses and procedures, must provide an easy search system
■ Is the basis for decision making and planning
Application Components: Medical Documentation System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
33
Application Components: Medical Documentation System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
34
A screenshot of a medical documentation system showing a list of patient-related documents on the left and an opened discharge letter on the right
■ A provider or physician order entry system (POE) (computer-supported POE systems are called CPOE) supports order entry.
■ This can comprise both order entry of diagnostic or therapeutic procedures as well as ordering of drugs.
■ POE systems support formulation of the order, appointment scheduling, printing of labels, and the communication of the order to the service unit
Application Components: Physician Order Entry System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
35
Application Components: Physician Order Entry System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
36
Screenshot of a lab order entry system. An order set of electrolyte has been chosen to see details (lower right). If available, results can already be reviewed (lower left).
■ Monitors, stores, and clearly presents a vast amount of patient-related clinical data in intensive care units (ICU)
■ May offer features for real-time decision-support and statistical analyses
■ Provide means for patient discharge and transfer to other wards or institutions, e.g. a short summary of the therapy on the ICU
Application Components: Patient Data Mangement System (PDMS)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
37
Application Components: Patient Data Mangement System (PDMS)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
38
Screenshot of a patient data management system showing a patient’s vital parameters and given drugs during a day
■ Supports operation planning and operation documentation as a specialization of execution of diagnostic and therapeutic procedures
■ It allows assigning of operation date and time, and therefore should be available on the wards as well as in the offices and management units of the operating rooms
■ During each operation, a vast amount of data has to be documented, including the members of the operating team, the operative procedure, the date and time, duration of the operation, materials (e.g. implants) used, and other necessary data to describe the operation and its results
Application Components: Operation Management System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
39
Application Components: Operation Management System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
40The operation management system shows the surgeons as well as thepatients that are assigned to them. For each patient, the planned procedure and the status of the operation is displayed.
■ The enterprise resource planning system (often: ERP system) enables a hospital to manage its financial, human and material resources
■ One major goal is the documentation and billing of all accountable services.
■ It thus supports hospital functions such as:
□ controlling,
□ financial accounting,□ facility management,□ human resources management,□ quality management and□ supply and disposal management
Application Components: Enterprise Resource Planning System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
41
■ Contains data which have been extracted from other application components
■ The data have been aggregated and transferred into a suitable format and then actively loaded into the data warehouse (push-principle)
■ Can be used for hospital management or for research
■ Sample questions: What are the top ten diagnoses of the patients treated in our hospital? Which department of the hospital causes the highest material costs? In which department do patients stay longest on average?
■ Analyzing these data by data mining techniques
■ A data warehouse system supporting hospital management is often called business intelligence system
Note: There is an ISO standard (ISO TR 22221) on “Good principles and practices for a clinical data warehouse
(CDW)” which provides implementation guidance for data warehouses in HIS
Data Warehouse System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
42
Data Warehouse System
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
43Screenshot of a data warehouse system. The diagram shows a target/actual comparison for the bed occupation of a university hospital
■ Retains all data independent of whether it will be used someday
■ Also unstructured data like sensor data, text and images are stored without transformation
■ Are designed for low-cost storage and fast ingestion of new types of data
Data Lake
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
44
https://www.searchtechnologies.com/blog/search-data-lake-with-big-data
■ HL7
■ FIHR
■ DICOM
■ ISO/IEEE 11073
■ CCOW
■ IHE
■ EDIFACT
■ CDA
■ …
Medical Data FormatsStandards for Data Exchange
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
45
■ Describes message and event types exchanged between application components
■ Event-driven communication
■ Example messages:
□ ADT (Admission Discharge Transfer)
□ ORM (Order entry message)
Health Level 7 (HL7)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
46
Message
Event
Acknowledgement
■ ADT (Admission Discharge Transfer)
■ 51 different types of ADT messages
Health Level 7 (HL7)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
47https://corepointhealth.com/resource-center/hl7-resources/hl7-adt
■ Builds on HL7 version 2.x and HL7 version 3.x
■ Supports both XML and JSON (RESTful API)
■ Building blocks: Resources, References and Profiles
■ Reasoning: 80/20 information model
Fast Healthcare Interoperability Resources (FHIR)
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
48https://www.hl7.org/fhir/
■ Building blocks: archetypes (define meaning), templates (assemble packages of data)
■ AQL – language for querying into the data
■ Clinical Knowledge Manager – online archetype manager
■ Reasoning: Maximal information model
openEHR
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
49https://www.openehr.org/
More on the Use of openEHR and FHIR
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
50
https://www.youtube.com/watch?v=A5MujeGGeqY
https://www.youtube.com/watch?v=QnRgQ_dvkqc
Standards for Data Exchange: IHE
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
51Source: http://www.ihe.net/
■ Specifications, tools and services for interoperability
■ Sharing of health information across systems
■ Coordinated use of established standards (DICOM, HL7)
■ Addresses clinical needs described in domains such as pathology or eye care
■ Targets care provision workflows in profiles describing actors, their transactions and messaging formats
■ Operational Data Model (ODM)-XML
■ Enables exchange of clinical study data
■ Models a Case Report Form (CRF)
Data Formats for Clinical Studies: CDISC ODM
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
52https://www.cdisc.org/
Clinical Data Interchange Standards Consortium
LIM
S W
iki:
Bas
ics
of c
ase
repo
rt f
orm
des
igni
ng in
clin
ical
res
earc
h
■ Operational Data Model (ODM)-XML
■ Enables exchange of clinical study data
■ Models a Case Report Form (CRF)
Data Formats for Clinical Studies: CDISC ODM
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
53https://www.cdisc.org/
Clinical Data Interchange Standards Consortium
http
://w
ww
.xm
l4ph
arm
a.co
m/C
DIS
C/im
ages
/CD
ISC_
OD
M1
1_d
raft
_exa
mpl
e.gi
f
■ Digital Imaging and Communications in Medicine
■ Stored in a Picture Archiving and Communication System (PACS)
Standards for Data Exchange: DICOM
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
54http://www.santesoft.com/win/sante-dicom-viewer-free/img/modalities.png
■ Number of databases being used to store data
■ Number of application components
■ Number of different software products and vendors
■ Patterns of communication
■ Types of integration
Architectual Characteristics
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
55
■ DB1:
□ often called the central database, one DB that stores all data
□ Schema and methods for access/storage of data must be known
□ Application components typically stem from the same vendor
■ DBn:
□ Several application components have their own
databases
□ Data is stored redundantly
Number of Databases: Central vs. Distributed
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
56
Number of Application Components: Monolithic vs. Modular
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
57
■ AC1:
□ One application component supporting all hospital functions
□ Usually not sufficient, typically specialized software
■ ACn
□ Modular architecture with more than one application component
□ Applications might share a single database
DB1, AC1 architecture
(DB1, ACn) architecture with multiple computer-based application components
Number of software products and vendors: all-in-one vs best-off-breed
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
58
■ V1:
□ One vendor, homogenous all-in-one architecture
□ Independent from AC
■ Vn:
□ Several vendors, heterogeneous best-of-breed architecture
□ Typically DBn
Communication Pattern: Spaghetti vs Star
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
59
■ Application components being part of an ACn architecture have to be connected to achieve integration
■ Spaghetti Communication Pattern, CPn :
□ Directly connect to ACs that need to exchange data
□ Increasing number of bidirectional
communication interfaces
(DBn, ACn, Vn, CPn) architecture
Communication Pattern: Spaghetti vs Star
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
60
■ Star Communication Pattern, CP1:
□ Most DBn systems use a communication server as middleware
□ n interfaces for n applications
(DBn, ACn, Vn, CP1)
Integrity and Integration
61
■ Object Identity: Patient gets a unique PIN, for each case a CIN
■ Referential Integrity: Correct assignment of CIN to patient or results to case
■ Consistency: Redundant data in DBn systems must be kept identical
■ Data Integration: Data is available wherever needed, no double recording/change
■ Semantic Integration: Sender and receiver interpret data in the same way,
e.g. ICD-10 for diagnoses, SNOMED for medical terms, LOINC for lab procedures
■ Access Integration: Applications are accessible where they are needed
■ Presentation Integration: Applications present data in the same way
■ Contextual Integration: e.g. non-repeated sign-on and patient selection
■ Functional Integration: Functions used by several applications are implemented once
■ Process Integration: uses all other integration types to conform to business processes
1. Patient admission event results in “A01“ HL7 event
2. Arrange data about patient and case in a HL7 message
3. Sending of message to communication server
4. Multicasting of message to all applications that need it
5. Applications receive message and store data in their DB systems
6. When updating patient/case over master application, “A08“ HL7 event is triggered
7-9. Arrange updated data in a HL7 message, send it to communication server and multicast to respective applications
10. All applications receive message and update DBs, data integrity is ensured.
Communication Server Example in a (DBn, ACn) Environment
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
62
Medical Data Integration Center to leverage data between hospitals
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
63
Medical Data Integration Center Components
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
64
HiGHmed–an open platform approach to enhance care and research acrossinstitutional boundaries, Haarbrandt et. Al, 2018
Gesundheitscloud Architecture
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
65
■ Storage of encrypted personal health data
■ Display your own data
■ Provide access to your data
■ User apps on top e.g. for analytics
■ Research platform leverages de-identified
data
What to Take Home?
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
66
■ Medical informatics is a broad field with many different standards, terminologies and applications
■ There is an abundance of data and processes in hospitals
■ You know about the challenges involved in transforming clinical needs into IT solutions
Chapter 6
Reading Materials
Data Management forDigital Health, Winter term 2018/19
Clinical Software Architectures
67