Evolution of Data Management System for SARS Outbreak

1
Evolution of Data Management System for SARS
Outbreak
IT in Health Forum

• Dr Heston Kwong
• Principal Medical Officer
• Department of Health
• 30 July 2003

2
Acknowlegement

• Hong Kong Police Force
• Hospital Authority

3
Outline of Presentation

• Principles of data management
• Overview
• Enhanced data management system
• Significant impacts
• Future challenge

4
Data Management

• Data capture
• Problems detection
• Outbreak monitoring
• Data analysis
• Information extraction
• Alerts communication
• Response formulation

5
Overview of Notification System
6
Statutory Notification System

• Statutory notification system (Cap. 141)
• Statutory notification form, through fax and
  phone

7

• Face to face interviews and field visits by field
  epidemiologists

8
EPI-INFO v.6 for storage and analysis
9
SARS Outbreak

• 1755 SARS cases
• March to June

10
SARS Outbreak

• Unprecedented outbreak
• Multiple foci
• Requires timely response demands immediate
  public health measures
• Traditional data management system not sufficient
• No data and network sharing between DH and HA
  systems
• SHWF initiated online computer system to access
  patient information directly by DH

11
Enhanced Data Management System

• eSARS
• MIIDSS
• SARS-CCIS

12
eSARS
Network Communication
13
Hospitals and Clinics in Hong Kong

• 43 Public Hospitals
• 46 Specialist Outpatient Clinics
• 64 general outpatient clinics
• 12 private hospitals
• HA has an internal network to connect public
  hospitals

14
DH Broadband Secure Private Network
IP/VPN network
15
Network connection between HA and DH
16
eSARS

• Two components
• Case list component
• Contact tracing component
• Case list component hospital admission, clinical
  information, clinical status
• Contact component enable Designated Medical
  Centres conduct 10-day medical surveillance,
  capture information of close contacts
• Launched on 12 April 2003
• Information downloaded to DH

17
(No Transcript)
18
(No Transcript)
19
MIIDSS

• Requires a tool quickly identify linkage between
  cases and contacts
• Purpose for rapid implementation of public health
  measures
• MIIDSS-crime investigation programme, designed to
  identify linkage between people, between people
  and events
• Police produce regular hotspot on potential
  clustering
• Hotspot reports commenced on 13 April 2003

20
Cluster Relationships
Overview of the clustering of SARS cases in HK
An individual cluster of SARS cases
The largest cluster of SARS cases identified
21
Cluster Relationships
LAU XX Cluster onset 12 March (The first
private medical practitioner died of SARS)

• Among 6 cases of this cluster, the earliest onset
  date was 12 Mar of a 13-year-old girl (YIP) who
  visited Dr Laus clinic on 12 Mar.
• Another case, a 3-year-old boy (CHU) visited Dr
  Laus clinic on 20 Mar.
• Onset date of Dr LAU was 20 Mar and his wife got
  infected with the onset on 26 Mar.

22
Identify high risk locations
To maintain an overview of the overall
geographical distribution of SARS cases in Hong
Kong
23
Identify high risk locations- Case study of Sau
Mau Ping district

• By the league table, MIIDSS facilitates the
  tracking of the accumulation of SARS cases in
  each district / housing estate / building to flag
  up high risk locations of SARS occurrence.

The league table indicated that Sau Mau Ping
(SMP) district had the highest number of
accumulative SARS cases on 20 April 2003, most of
which came from Amoy Gardens.
24
Identify high risk locations- Case study of Sau
Mau Ping district
25
Identify high risk locations- Case study of Tai
Po district

• MIIDSS league table of 20 April indicated that
  Tai Po district had the second highest number of
  accumulative SARS cases after Sau Mau Ping at
  that time.

26
Identify high risk locations- Case study of Tai
Po district
27
Spread pattern analysis
A noticeable spread pattern of SARS on
consecutive floors of the same vertical stack of
units, particularly units 07 and 08 Spread
pattern was also tracked by the onset dates and
address locations of the SARS cases in Block E,
Amoy Gardens with a colour scheme to indicate the
sequence of onset dates (from darker colours to
lighter ones). Investigation efforts were
promptly steered to look into the environmental
factors possibly transmitting the disease along
the same vertical stacks of buildings in Amoy
Gardens.
28
Spread pattern analysis
Continuous monitoring of the latest spread
pattern
Accumulative SARS cases
29
Patients categorization
Infected Persons by Category
30
MIIDSS targets atResponse Management
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
SARS-CCIS

• EPIINFO v.6 not run on shared network
• Few opportunities to efficiently consolidate and
  compare information kept at separate dataset at
  Regional Offices
• Enhance existing data management system to
  provide common dataset and common questionnaire
  in electronic format
• Enable construction of cluster trees
• Launched on 2 May 2003

37
(No Transcript)
38
(No Transcript)
39
Data Analysis
40
(No Transcript)
41
Data Analysis
Time
42
Data Analysis
Place

• Geographical information extracted from central
  dataset
• Produce hotspot report
• Identify potential clustering

43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
Data Analysis
Person

• Need to identify linked up cases and contacts in
  cluster
• Cluster tree
• Cluster records for epidemiological analysis

S0300963
YOUNG XXX XXX
YXX's nearby
patients
S0301312
S0301217
S0301539
S0301567
LAW XX XXX
WONG XXX XXX
CHUI XXX XXX
LEE XXX XXX
WXX's nearby
WSF's nearby
patients
patients' visitor
S0301586
S0301754
S0301775
S0301710
CHAN XXX XXX
LAM XXX XXX
WONG XXX XXX
TANG XXX XXX
LXX's hcw
LXX's nearby patients
S0301744
S0301755
S0301760
S0301770
S0301798
S0301804
MOK XXX XXX
LI XXX XXX
CHAN XXX XXX
CHUI XXX XXX
FUNG XXX XXX
POON XXX XXX
WTS PXX's nearby
patients
S0301799
S0301803
LAI XXX XXX
LAM XXXX XXX
48
Cluster Tree
49
Cluster Tree
50
The SARS Outbreak
51
Epidemic Curve
52
Number of affected healthcare workers
53
Proportion of Occupational Groups
54
Notification of SARS

• SARS diagnosed
• Doctors fill up electronic medical record
• Immediately accessed by field epidemiologists at
  Department of Health

eSARS
55
Contact Tracing

• Contact data collected through interview,
• Updated central dataset through online computer
  screen
• Field epidemiologist can review and update
  information collected by other field
  epidemiologist at other location

SARS-CCIS
56
Police
HA
Clinical information transmitted to eSARS
Cases admitted to SARS cohort wards in Hospital
Authority
SARS CCIS
57
(No Transcript)
58
(No Transcript)
59
MIIDSS hotspot report
eSARS database
SARS-CCIS interface
eSARS web interface
INSPIRE Web interface
IP/VPN internal private network
SARS-CCIS (database)
60
Significant Impacts (1)

• Timely access of patient information
• Instantaneous updating of confirmed and suspected
  cases
• Saved at least one day between identification of
  cases in wards and notification to DH

Saved at least one day
61
Significant Impacts (2)

• Hotspots report from MIIDSS facilitate early
  identification of potential clustering of
  confirmed and suspected cases
• Enables to plan field visits and proactively
  deploy resources to handle potential problems

early identification of potential clustering
62
Significant Impacts (3)

• Real time dataset enable assessment of outbreak
  development in terms of time, place and person
• Daily situation report on epidemic curve and
  summary of public health measures response
  formulation
• Timely generation of action lists alerts
  communication

time,
place and person
response formulation
alerts
communication
63
Significant Impacts (4)

• Updated statistics and outbreak situation
  reported in SARS bulletin and published on SARS
  website
• SARS affected building published on SARS website

SARS bulletin
SARS affected building
64
Data Dissemination

• Effective data dissemination
• Shorten time to initiate investigation
• Facilitate other departmental teams assist to
  implement public health measures
• Inform public on outbreak situation

65
Communication public education

• gt 89,000 calls answered by telephone hotline

66
Communication public education

• Daily media briefings

67
Data Dissemination
SARS-CCIS Central dataset
SARS bulletin
Building lists of SARS cases
Regular press conference
68
Future Challenge

• Establishment of data exchange network
• Data alignment, standardisation
• Data dictionary
• XML Schema and repository
• Intelligence exchange among cities and countries
  in neighbourhood
• Real time detection alert system
• Syndromic surveillance
• Mathematical modelling and data mining

69
Data Management

• Data capture
• Problems detection
• Outbreak monitoring
• Data analysis
• Information extraction
• Alerts communication
• Response formulation

70
Thank You
71
Data Analysis
72
Data Dissemination
73
Lessons Learnt

• Timely access of epidemiological and clinical
  information
• Geographical analysis to identify cluster
• Cluster analysis to identify risk factors
• Central dataset and use of IT

74
Thank You