STRATEGIES FOR BUILDING NATIONALSCALE LONGITUDINAL ELECTRONIC PATIENT MONITORING SYSTEMS FOR HIV TRE

1
STRATEGIES FOR BUILDING NATIONAL-SCALE
LONGITUDINAL ELECTRONIC PATIENT MONITORING
SYSTEMS FOR HIV TREATMENT AND CARE IN PEPFAR
COUNTRIES

• October 25, 2007 ? Lusaka, Zambia

2
Botswana Integration at the Top

• Suzanne Cloutier, MSPH
• Informatics Consultant BOTUSA

2
3
Overview

• Current HMIS Environment
• IT Infrastructure
• GoB Organizational Structure
• HIV Data Warehouse
• Challenges and Opportunities
• Next Steps

4
Current HMIS Environment

• Botswana has several information systems for
  capturing data on health services provided to
  HIV patients
• EMR Integrated Patient Management System (IPMS)
• ART Patient Information Management System
  (PIMS)
• TB Electronic TB Register (ETR), Isoniazid
  Preventative Therapy (IPT) System
• VCT VCT Management Information System
• General District Health Information System
  (DHIS)

5
Current HMIS Environment

• Integrated Patient Management System (IPMS)
• Proprietary, centralized, client/server
  application implemented by Meditech in 2003
• Focused on overall patient care and treatment in
  community, clinic, and hospital settings
• Provide user-friendly, modular, scalable
  application
• Focus on core functions
• Phase 1 4 hospitals 16 satellite clinics
• Accounts for 48 of ARV patients

5
6
IPMS - continued

• Still in phase 1 pending evaluation
• Preliminary results from evaluation
• Inadequate user training
• Inadequate user support
• Inadequate bandwidth on the network
• No interfaces to other GoB systems
• Very expensive
• More functionality ? Higher cost

6
7
Current HMIS Environment

• Patient Information Management System (PIMS)
• Decentralized facility-based application
• MS Access backend supported by VB and SQL code
• Clinic data are entered by data entry clerks
• Standardized reports available for local and
  national use
• Installed at 28 ART sites not connected to IPMS
• Accounts for 52 of ART patients
• Upgrade necessary to support future needs

7
8
Current HMIS Environment

• TB treatment and prophylaxis
• Electronic TB Register (ETR) and Isoniazid
  Preventative Therapy (IPT) System
• Both developed on VB.Net platform with a MS
  Access backend
• Distributed systems with health facility data
  entered at district level and dispatched to a
  central database at the national program
• The TB systems are not integrated
• Ideally they would share district, facility, and
  patient tables

9
Current HMIS Environment

• Tebelopele VCT Management Information System
  (MIS)
• Decentralized application
• Developed in MS Access
• Installed at 16 sites across Botswana
• Counselors use handheld PCs to capture
  confidential client data
• Data collected at the VCT centers are converted
  to an Epi6 dataset, then merged into 1 dataset at
  the national level

10
Current HMIS Environment

• District Health Information System (DHIS)
• Developed in MS Access
• Uses Excel for analysis pivot tables, graphs
• Supports XML
• Installed as a stand-alone system in the
  districts
• Aggregated health facility data are entered at
  district level by Community Health Nurses
• Data entered support minimal indicator set
  defined by stakeholders
• Piloted in 4 districts roll-out to all
  districts will start before end of year

10
11
IT Infrastructure

• Government Data Network (GDN)
• TCP/IP
• connectivity
• via LANs
• Email
• Internet
• Web hosting
• Applications

11
12
IT Infrastructure

• Health Districts
• Comprehensive IT assessment of all health
  districts last year
• Conducted site visits to 24 districts and 3
  sub-districts
• Assessed the ICT infrastructure, as well as the
  capacity of key district personnel to use IT
  resources effectively and efficiently for
  managing strategic information
• Assessment teams consisted of 1 IT specialist and
  1 Informatics specialist

12
13
Health District Informatics Assessment

• ICT Infrastructure
• 135 computers were assessed
• Operating system
• Processor speed
• Hard disk size
• RAM
• Network connectivity
• Software
• Other ICT equipment

13
14
Health District Informatics Assessment

• Use of ICT Resources
• 122 district health officers were interviewed
• Date management activities
• File management skills
• Software use and skill level

14
15
GoB Organizational Structure
15
16
GoB Organizational Structure
16
17
GoB Organizational Structure
17
18
GoB Organizational Structure
18
19
HIV Data Warehouse

• Rationale for this Solution
• Masa ARV program decided to develop a data
  warehouse to integrate ARV data from IPMS and
  PIMS
• Agreed later to expand the scope to include all
  patient level data from the various health
  programs

20
HIV Data Warehouse

• Rationale for this Solution
• Fragmented approach to patient management and
  outcome monitoring
• Each program captures data independently and is
  unable to share data with other programs
• Level of detail for data varies across programs
• ARV, TB, and VCT programs capture patient level
  data
• Other health data are aggregated by facility
• Applications are built on different platforms
• No standard way to exchange data
• Not easy to link patient records within programs,
  let alone across programs

21
HIV Data Warehouse

• Rationale for this Solution
• Not possible to get a comprehensive picture of
  the services PLWHA access, nor the overall
  effectiveness of the healthcare provided to them
• Goal of the HIV Data Warehouse
• Integrate all electronic health data on services
  provided to PLWHA so that analysis and reporting
  may be done to improve patient care and outcomes

22
HIV Data Warehouse

• Objectives
• Integrate ARV data into a national data
  repository
• Integrate other health services data for HIV
  patients
• Allow access to anonymous, linked health data
• Ensure appropriate privacy/confidentiality and
  security measures are implemented

23
Phase 1 Integrate ARV Data
HIV Data Warehouse
24
Phase 2 Integrate TB Data
HIV Data Warehouse
25
Phase 3 Integrate Other Data
26
HIV Data Warehouse

• Progress to date
• Purchased server hardware and software
• Quad dual-core processors
• MS SQL Server 2005
• Defined the scope
• Data extraction, cleaning, integration,
  transformation, loading, reporting, business
  intelligence component
• Defined the development methodology
• Kimball methodology for MS SQL Server 2005 -
  The Microsoft Data Warehouse Toolkit by Mundy
  et al.

27
HIV Data Warehouse

• Progress to date
• Gathered user requirements from ARV program
• Developed the initial data model
• Received TA from WHO data warehouse expert
• Extracted and loaded IPMS data onto the server
• Able to produce monthly ME reports
• Provide limited research support
• Started development of prototype based on
  pharmacy module as proof of concept

28
Preliminary Data Model
Fact Tables
Care Summary
Conformed Dimensions
Dimensions
Observations
Observation
Person
Lab Results
Lab Test
Facility
Drug Orders
Encounter
Drug
Drug Stops
Transfers
29
Privacy Protection

• Support for privacy and confidentiality is built
  into the data model
• Demographic data will be stored separately from
  the data warehouse
• Pseudonymous patient identifiers will be used
  within the data warehouse
• Overall design allows for the implementation of
  basic privacy principles through engineering and
  policy

30
Security

• Foundation for privacy protection
• Security will be layered into the data warehouse
  system using industry standard approaches
• Data model includes components to ensure logging
  and auditing capabilities

31
Preliminary Data Structure
Fact Tables
Personal Data
Care Summary
Identifiers
Conformed Dimensions
Dimensions
Private Details
Observations
Observation
Person
Lab Results
Lab Test
Facility
Private Contacts
Drug Orders
Encounter
Drug
Drug Stops
Audit and Tracking
HIV Data Warehouse
Transfers
32
ARV Enrollment Projections
125,000
110,000
95,000
75,000
33
Estimated Volume of ARV Data
33
34
Challenges

• Non-technical
• Low level of communication and collaboration
  across government organizations
• The GoB makes all decisions
• Nothing is implemented without their approval
• They operate on their own timetable with their
  own protocol
• No CIO to provide strategic vision, sponsorship,
  or guidance to the project team
• No national health informatics strategy

34
35
Challenges

• Technical
• Shortage of skilled IT resources
• Much of the patient data are in paper-based
  systems
• Modeling atomic drugs vs. combination drugs vs.
  drug regimens
• Network Infrastructure
• Very few healthcare facilities are connected to
  LANs
• GDN is not robust or stable enough to support
  centralized systems

35
36
Challenges

• Technical
• Remote sites
• Minimal connectivity via internet or cellular
  network
• Synchronization issues
• Data transfer challenges
• Linking patient records from disparate systems
• Patients records are fragmented
• Data integrity problems
• Misuse of national ID (Omang )
• Patient names

36
37
Opportunities

• National HMIS Strategic Plan Development
• Funds requested in COP 08
• Provide guidance for IS development
• Include international ICT standards and
  guidelines
• Active participation by 5 key GoB stakeholders
• Revitalization of a System Integration Working
  Group in DHAPC

37
38
Opportunities

• DIT E-Governance Initiative
• by December 2010
• All healthcare facilities in the country will be
  networked
• Healthcare information and applications will be
  made available online
• Pilot Implementation of Security Guidelines
• MS SQL Server Security Consultant
• Set up security measures per international
  guidelines

38
39
Next Steps

• Complete Prototype of Pharmacy Module
• Evaluate modeling of
• Combination drugs (e.g. Combivir) vs.atomic drugs
  (e.g. AZT, 3TC)
• Drugs vs.drug regimens (e.g. DDID4TEFV)
• Evaluate MS SQL Server analysis tools
• Acquire and install business intelligence (BI)
  tools

39
40
Next Steps

• Set up security for MS SQL Server 2005
• Provide on-the-job training to local IT staff
• Use probabilistic matching software to link
  patient records
• Evaluate COTS product
• Further refine data extracts from IPMS to include
  hospitalizations
• Extract and load PIMS data

40
41
Thank You!
Thank You!
Thank You!
Thank You!
Thank You!
41
42
Background

• HIV/AIDS
• Prevalence is declining, but still high
• 17 in general population (2004)
• 25 in adults 15-49 years
• 32 in pregnant women 15-49 years (2006)
• PMTCT program began in 1999
• Full coverage since 2003 with uptake of about 96
  
• 96 of babies born to mothers in PMTCT are HIV-
• ARV program began in 2002
• Out of 1.7 million population, approx 270,000 are
  PLWHA
• gt82,000 of the eligible 91,000 PLWHA are on ART

42
43
Background

• TB
• Case Reporting
• 10,000-12,000 /year
• 60-86 of TB patients are HIV
• Isoniazid Prevention Therapy (IPT) program began
  in 2002
• 56,000 PLWHA screened and gt42,000 enrolled as of
  2006

43
44
Progress with ARV Data

• Extract monthly IPMS data to produce tables 27
  of the Site Managers Report
• Extract 1.5 million pharmacy records of ARV drugs
  dispensed
• Extract 8 million lab records of lab tests done
  for patients on ARV
• Currently integrating MASA data from the main
  hospital sites

45
BRHIMS OVERVIEW
Data aggregated by DAC sent to MOH / MLG
MOH
MLG
NATIONAL LEVEL
24 health districts with DHIS(2 yrs operational)
DAC
DHT
27 LG districts
DISTRICT LEVEL
Forms-based data collection
OVC, District admin
Clinics
LEGEND Data flow MLG Ministry of Local
GovtMOH Ministry of HealthDHT District Health
TeamDHIS District Health ISDAC District AIDS
Coordinator