Transcript Applied Informatics

Using the NII to
Coordinate Health Care
1. Applied Informatics
2. WebCIS
3. Home PFT
Applied Informatics
George Hripcsak
Nilesh Jain
Charles Knirsch
Ariel Pablos-Mendez
What

overall
 use
the National Information Infrastructure
(NII) to coordinate care for patients across
multiple encounters, providers, and settings
 clinical area
 begin with treatment of tuberculosis
 then extend technology to all patients
Who

Columbia-Presbyterian Medical Center
 out-

and inpatient
New York City Department of Health
 tuberculosis

clinics
Visiting Nurse Service of New York
 patient’s
homes
Why - goals
coordinate (TB) care among providers
 respond to patient needs
 reduce variance in care via TB protocols

 reduce

treatment failure, resistance, spread
demonstrate privacy and security
How - components
connectivity
 linking electronic medical records
 automated clinical protocols
 information resources
 wireless pen-based computing
 security

Schematic
registration
CPMC
VNS
TB isolation
visit data
case
reporting
DOH
clinical
data
DOH
kiosk clinics
kiosk
mobile computing
mobile
computing
home
Connectivity
Internet
 CPMC-VNS via linked frame relay
 CPMC-DOH via dial-up phone line

TB patient resources

patient education critical in TB
 patients

do not have Web access
clinic kiosk
 Web
browser
 touch screen monitor
 HTML with large buttons, discrete screens
 statistics gathering
TB patient resources

patient information
4
TB pamphlets, 2 in Spanish
 62 English pages, 31 Spanish pages
TB patient resources
...cuando las defensas del cuerpo estan debiles, las bacterias inactivas de la
tuberculosis se reactivan y se salen de las paredes
TB patient resources

utilization (8/95-1/96)
 275
pages (44 doc) per clinic day
 40% Spanish
 (100 repeat visits, 2 new visits per day)

clinic director
 addressed
language barriers
 patients asked better questions
 only once personnel encouraged patients
TB Web resources

on Internet
 patient
pamphlets
 DOH’s TB protocols (100 pages)
 links to other sites

utilization
 25,000
files per month
 2000 unique computers
 12% outside US
TB resources
well-used
 need to prove kiosk effect
 how to address clinical questions from
Web users

Mobile computing

home care nurses
 isolation
 patient
information and changes
 carry manuals

use wirless mobile computing
Mobile computing

pen-based tablet (Fujitsu)
 2.5

lbs, 50 MHz 486, 170 MB disk
CDPD wireless telecommunications
 90%

connection, rest store and forward
applications
 work
lists, initial visit
 data forwarded to CPMC
 information resources (care plan, policies)
Mobile computing

8 nurses for 3 months
 enthusiastic
 increase
in information (contacts)
 less need to carry manuals
 empowerment (contact with CPMC nurses)

but no paperwork reduction
 did

not automate everything
coordination: MDs do not have devices
TB detection & reporting
automatically report CPMC tuberculosis
cases to DOH
 clinical event monitor

 countable:
TB culture
 suspicious: AFB smear, lab tests, CXR,
medications

natural language processing
TB detection & reporting
improved timeliness (2 weeks)
 could not automate entirely

 lack
of electronic clinical information
(clinical improvement, PPD)
 difficulty automating complex judgments
(lab errors)
TB isolation

4% of new TB pts infected in hospital
 respiratory
isolation
 surveillance and enforcement is difficult

automated detection of patients
 at
high risk for TB
 not in isolation room
TB isolation

alerts based on electronic patient data
 “The
patient's chest X-ray (on 12 Oct 1995
at 12:11) shows specific evidence for
tuberculosis disease. The patient is in the
hospital, NOT in an isolation room.”
 alerts are sent to
 hospital epidemiologist
 clinician (via electronic medical record)
TB isolation
43 patients proven TB (7/95 to 7/96)
 13 (30%) not isolated by MD
 5 (38% of 13) caught by system
 2 of 30 taken off isolation too soon,
system recommended re-isolation
 15 FP for every 1 TP (PPV 6%)

TB discussion
only critical tasks are achieved
 largely intra-organizational gains
 security
 standards
 difficult to evaluate a diffuse project

TB discussion

cost benefit
 TB
detection & isolation save $10,000/year
 only pays incremental costs
 entertainment, commerce drive Web
WebCIS
James Cimino
George Hripcsak
Soumitra Sengupta
Socrates Socratous
WebCIS
Web-based clinical user interface
 three-tiered architecture

 mainframe
–
TCPIP socket interface
 UNIX
–
Web server
CGIs in C
 Web
–
(DB2) clinical repository
browser
HTML and Java Script
WebCIS

Medical Entities Dictionary
 translation
of codes
 design of displays
–

based on MED classes and slots
security
 time
out
 back in history
 Secure ID cards
WebCIS

benefits of Web development
 quick
prototyping and development
 improved access
 easier deployment and maintanence
 multimedia
 hypertext links
 security
WebCIS

challenges of Web development
 CGIs
stateless
 moving target
 security
Home PFT
Joseph Finkelstein
George Hripcsak
Manny Cabrera
Home PFT
monitor asthma severity in patients’
homes
 current technology

 symptom
 peak
reports
flow
 poor predictive power and reliability
Home PFT

components
 portable
spirometer
 handheld computer with data entry
(or desktop with Web browser)
 wireless or landline communications
 clinical repository with decision support
 Web server
Home PFT

results (7 normal, 3 patients)
 able
to perform PFT
 able to run computer interface
 1 (land) to 8 (RAM) minute upload delay
 current equipment fragile
 clinical annecdotes
–
–
intervene for morning exacerbation
normal peak flow with poor terminal flow
Home PFT

benefits
 full
flow-volume curve (with FVC)
 portable
 can check compliance
 immediately available to physician
 automated decision support
Home PFT

questions
 what
parameters best predict exacerbation
and is it preventable
 optimize user interface and communication
 what can be automated
Overall conclusions
enormous potential
 not just the Web

 clinical
repository
 automated decision support
 vocabulary tools
 kiosks, wireless

will be driven by other forces