Transcript Medical Informatics - Molde University College

Medical Informatics
For a Hospital Medical System
Setting
By Dr. James G. Williams
Medical Informatics
• Medical informatics is a broad field spanning electronic medical
records, telemedicine, information retrieval, image processing and
analysis, bioinformatics, and evaluation methodologies.
• Outcomes research is the study of the effectiveness of health care
using large data sources and other advanced technologies.
• Medical Informatics focuses on Health Care, including electronic
medical records, information retrieval, medical decision-making,
telemedicine, clinician information needs, artificial intelligence, and
outcomes research.
• As a result of medical informatics training, graduates are better able
to take on additional IT responsibilities in their existing careers or
embark on new careers as developers and managers of healthcare
IT systems.
• But what level of technical training is necessary to make a
graduate’s skill set adequate for the career opportunities?
Islands of Information in Hospitals
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Medical History
Order Processing
Blood Bank
Laboratory Tests
Pharmacy
ADT (Admissions, Discharge, Transfer)
Vital Signs
Transplant Units
Physical Therapy
Dietary
ICUs
Radiology, Cardiology, Nuclear Medicine
Out-Patient Clinics
Surgery
Billing Insurance, HMO, and Government
Scheduling Residents, rooms, equipment
Research Support
Best of Breed Approach
• Each functional area (department or lab)
wants the best system for their area.
• For example, SunQuest for Laboratory
Testing, MediPac for ADT.
• This creates multiple systems that are
unable to exchange information easily
• The task of the IT department is to
integrate the data from these systems
Integrated Medical Information
• Patient Medical Record should contain
every piece of relevant medical data ever
generated concerning a patient
• Patient Medical record should be available
to health care workers anywhere, anytime.
• Need to get the data into a digital form.
• Need to create a database capable of
storing data in multimedia formats.
Integrated Medical Information
• Need to provide access to only that data as
needed by the health care worker.
• Need to provide a means to correlate data from
various sources, e.g. symptoms, tests,
diagnoses, family history.
• Need to present data in a form and format that is
understandable
• Need to be Health Care Worker Oriented, not
function or department oriented.
• Need to protect the patient’s right to privacy.
• Can the Patient carry his/her medical record with
him/her?
Billing
Blood Bank
Medical History
Pharmacy
Health Care
Workers
Radiology
Integrated
Medical record
Physical Therapy
ADT
Dietary
Surgery
Vital Signs
Transplant
ICUs
Cardiology
Out-Patient
Clinics
Nuclear Medicine
Medical Archive
(Data Mining Source)
Medical
Records
Medical Archive
Database
Medical
Research
Information Retrieval, Artificial Intelligence,
Medical Diagnosis Support, Survival Intelligence
Medical Archive Record System
(MARS)
• University of Pittsburgh Medical Center has data
feeds from all departments to a system called
MARS.
• Surrogates are sent for image data that is
maintained in separate databases.
• Each source (department) has a tagging system
for individual data elements within the records or
documents they send.
• MARS has meta data describing the data from
each source.
MARS
• The data is stored in text format since this
allows flexibility in searching
• All data about a patient is available at all
times.
• Acts as a data warehouse for research
• Reference to images (X-Rays, sonograms,
etc. are maintained)
• Full text of reports are maintained
To Do List
A feature that helps keep track of the
status of tasks
Prescription Writing
Reduce the time spent refilling
prescriptions by using the list of
pre-built Rx codes or using previous
prescriptions made.
Patient Summaries and H
& P's
Track all of the pertinent historical
information on your patients.
Note generation
Quickly document a patient
encounter with pre-built templates
and codes.
All pertinent patient data
on one screen
Reduces the amount of time spent
searching a record for valuable
information.
Patient Pictures
Import a patient's picture into
demographics section of their
chart.
Free-text or template
driven entry
You can enter the patient data the
way you want to.
Lab/Radiology/Pathology
Reports
Store ancillary reports for quick
review.
All
Codes/Protocols/Reports
completely customizable
Generate patient information with a
look that's all your own.
Phone Message
Section
Record the who, what and when
of non clinic based encounter.
Vital Sign Tracking and
Charting
Tabulator format allows for
easy review of vital signs over
time and instant charting of
data.
Import Reports from
other systems
Import an ASCII or RTF
formatted report directly into
the chart.
Comprehensive
Demographics
Track all pertinent
demographics information
250+ Prebuilt
Protocols (Templates)
Use the included templates out
of the box or as a starting point
for your own documentation
guidelines.
Patient Education
Handouts
Make sure your patients get the
information they need by
providing easy to understand
informational handouts.
1000's of pre-built
Dx/Rx/Plan/Follow Up
Codes
Codes help to speed data entry
by replacing long blocks of
typing with small, easy to
remember codes that "expand"
to the full narrative.
Works with Voice
Recognition systems
Use the latest versions of
popular voice recognition
software to reduce the amount
of typing required.
Critical Design Factors
• Unique Patient Identification that will never
change
• Unique Identification for every health care
worker
• Unique Identification for every product,
e.g. blood product, X-ray, etc.
• Unique Identification for every service, e.g.
transfusion, physical therapy, etc.
AI in Medical Informatics
• Problem-Knowledge Couplers help to
guide the medical care of an individual by
soliciting information from a patient about
a medical condition, and then coupling, or
linking that information to a medical
knowledge database. The result of the
coupling is a list of care recommendations
based on the patient's unique set of
medical findings.
Medical Couplers
• Screening Couplers, such as Wellness, Physical
Examination, and Medical History, are used to
discover medical problems.
• Diagnostic Couplers, such as Chest Pain, and
Knee Problem, are used to help determine the
cause or causes of medical problems
• Management Couplers, such as Asthma,
Hypertension, and Diabetes, are used to
manage medical problems where the cause is
known.
Telemedicine
• Home telemedicine for the management of
diabetes in elderly patients in medically
underserved areas. Informatics research
areas include logistics, user interfaces,
patient education, user interfaces, security
and videoconferencing technology.
Discovering Knowledge
• Machine learning and visualization are
methods for discovering knowledge in
large databases. Research focuses on
testing and extending existing discovery
methods to improve their performance on
clinical data. Important issues include
training set size, data accuracy, data
completeness, and representation.
Patient Access to Medical Records
• Patients' access to their own medical
records is assumed to be a consequence
of the emergence of electronic medical
records and is being mandated by federal
regulations. Yet, very little is known about
how patients will fare in comprehending
and using the information, and how it will
impact the patient-caregiver relationship.
Reducing Medical Errors
• Two common causes of medical errors are
– inadequate access to information needed for
decision making and
– ineffective communication among members of the
patient care team.
– Reduce errors by (a) link patient data to relevant online information resources, (b) a virtual whiteboard to
facilitate the coordination of patient care tasks among
the house staff, attending physicians and nurses, and
(c) deployment on a wireless, handheld platform.
Medical Decision Making
• University of Utah Study has suggested
that nurses use different problemstructuring and knowledge-structuring
principles than those used by physicians
for medical decision making
Nursing Informatics
• (a) data and information are symbolic
representations of the phenomena with
which nursing is concerned,
• (b) expertise in problem structuring is
domain specific and
• (c) algorithms and heuristics used in
solving domain problems are peculiar to
nursing.
Critical Design Factors
• Interfaces must be tailored for the health
care worker, e.g. physician, nurse,
pharmacist, radiologist, lab tech, etc.
• Minimize keying and use scanning
technologies, e.g. bar codes, optical
character recognition, etc. to avoid
mistakes.
• Date and Time Stamp every piece of data
Critical Design Factors
• Allow for multimedia data objects, optical
(X-rays, sonograms, strip charts…), voice
(dictations), numeric (test results), text
(medical history, diagnosis, orders)
• Provide for indexing optical and voice data
objects either automatically or manually
(Content Analysis, Voice Recognition)
• Provide for searching across medical
records and finding patterns
Critical Design Factors
• Provide for drug interaction verification
• Provide high speed access and
presentation
• Provide for remote and mobile access to
archive and ICU monitoring data
• Utilize colors for presentation of test
results that indicate safe/negative (green),
caution/borderline (yellow),
positive/danger (red)
Critical Design Factors
• Provide standard statistical analysis
capabilities (BMDP)
• Provide graphing capabilities
• Provide for report writing templates from
medical test results (Phrase Paragraph)
• Provide for text to voice or voice to voice
over the telephone system of
test/diagnosis for referring physician
Critical Design Factors
• Use Visualization techniques for presentation of
massive amounts of data.
• Provide for on-line help for every data element,
e.g. test result, drug, conversions (milliliters to
liters, grams to milligrams, etc.)
• Track every access to data for security, privacy,
and utilization purposes.
• Fault Tolerant Systems
• Off the shelf system software – Not MUMPS
Visualization of Multiple Characteristics
(Female)
Sex - Male
W
e
i
g
h
t
Economic Status (Average)
BP (Average)
Age
High)
(BP High)
Technologies Required
•
•
•
•
•
•
•
•
•
High Speed Computers
Large Primary Memory
Massive and Fast Storage Devices (disk)
Graphical Devices (high resolution)
Digital X-Ray, Sonograms, etc.
High Speed Network
Mobile and Handheld interface devices
Scanning Technologies (Bar Code, OCR)
Voice Processing Hardware
Technologies Required
•
•
•
•
•
•
•
•
•
Database Management Systems
Graphical Software
IR Software
AI software
Voice Recognition and Text to Speech Software
Decision Support Software
Statistical Analysis Software
Application Specific Software
Fault Tolerant Systems