Information systems

1
Information systems Infrastructure
StandardsJørn Braa Information systems
Information flows processing computers,
databases/ applications networks, Intra-
Internet organisations peopleInformation
Infrastructure a larger compilation of the
above Many information systems messed together
(but working somehow)Complexity interaction
between many organisational IS structures that
are evolving changing the interactions are
therefore also changing Standards In one
perspective what makes it work But since the
world is changing standards must be easy to
change (or else constrain evolvement/
change) Flexible Standards
2
Information Systems

• Wide understanding
• from online booking systems, to electronic or
  paper based records systems for books or
  patients in hospital or in library
• AND the SOCIAL ORGANISATION OF IT!
• Networks, computers, software (from Operating
  Systems to Applications), organisational
  procedures and social organisation - are all
  part of an information system
• Technology is rooted in people (knowledge) not
  things organisation, knowledge, learning,
  innovation - as well as artefacts
• Information Systems are best understood as social
  systems
• a web of social and technical elements
• nothing in a technological system is poorly
  technical (non-human)
• the information system is part of the social
  context
• The history is part of the information system

3
Information Systems Information Infrastructure

• Large interacting Information systems may be
  regarded as information infrastructure
• Infrastructure has a wider meaning the technical
  grid (roads, wires, airports, etc.) as well as
  the institutions, support systems, organisations,
  schools, universities etc. necessary to run and
  maintain it.
• Information systems make up a web of more or less
  interacting information systems - more or less
  integrated through standards / lack of standards
• Web - network metaphor
• Develop over time - makes changes difficult
• Health Information Systems gt Health Information
  Infrastructure

4
Information Infrastructure Complexity

• Information infrastructures are complex
• A complex system is composed by multiple more or
  less interacting and interdependent (sub-)
  systems or components, which are developing
  following their own agendas and as responses to
  interaction with other components Therefore
• Uncertain and non-linear development not
  predictable
• Complex Adaptive Systems components/sub-systems
  are adapting based on feedback through
  interaction, thereby
• Causing emergent order, emerging order
  bottom-up without central control
• Actor Network Theory order created through
  alignment of network also bottom-up without
  central control

5

• Complexity infrastructure perspective
  Consequences for system development design
  Flexible standards
• Everything is part of something bigger - or rely
  upon something else
• Development needs to involve negotiations with
  many other systems
• Each system or world based on their own
  standards
• Since development is non-linear, within each part
  - for the wholeone global set of standards not
  possible
• Would constrain development
• Sub-systems are changing /adapting as response to
  interaction
• Communication/interaction between (Sub-)Systems
  components based on standards and clear
  interfaces
• Then Standards need to be flexible!
• Gateways ad-hoc ways to establish interfaces/
  translations between components

6

• Standards needed for interaction. But what when
  each component / system is gradually changing?
• How to ensure flexibility?
• Strategy, when designing information
  infrastructure
• Start with limited and targeted usefulness must
  know rather than nice to know
• Draw upon what is already there the installed
  base
• Build further on what is working and as learning
  take place
• Make Information Infrastructure design simple
• Modularise by building seperately key functions
  for each infrastructure through
• Layering distinguish between technical layers
  and data layers (e.g. dont lock particular
  information needs to particular software
  application), and
• Gateways rather focus on appropriate linkages
  between sub-systems than building one big system

7
(Sub-) Systems interacting through standards and
interfaces
System A
database
database
database
report
Shared Standards
database
database
report
database
database
database
database
report
database
System C
System B
Gateway ad-hoc Flexible way of establishing
interface compatible standards
8
DNHDP Western Cape
City Health
Births Deaths Notifiable diseases
New /emerging flow of information
City Health Clinic 3
City Health Clinic 1
Groote Schuur Hospital
Outside hospitals
City Health Clinic 2
City Health Clinic 4
School Health
Geriatric Services
PAWC
MOU (Midwife obstetric unit) PAWC
Day Hospital DNHPD
City Health Clinic 5
DNHDP Pretoria
Private hospital 31 medical specialists
Psyciatric hospital PAWC
Environmental office
Dental unit 1 PAWC
54 private medical pract.
RSC
Dental unit 2 PAWC
23 private dental pract.
UWC Oral Health Centre
12 private pharmacies
Dental unit 3 PAWC
12-15 NGOs
Mandalay Mobile clinic RSC
Youth Health Services
MITCHELLS PLAIN DISTRICT
South Africa after Apartheid fragmented and top
down health structure Also Districts very
fragmented
9
Multiple flows of information, reporting
structures, procedures paper and computer based
tools etc. acting people social structures -
Illustration of Information Infrastructures in a
district

A district
District management team
Local Government
District manager
Manager
District hospital
Environmental Health
district database
Register
Information officer
report
Information officer
Nurse
Immunisation program (EPI)
report
Information responsible
Nurse
Patients
Health centre/ Clinic
School Health
NGO
Register
Action
Clinic
Manager
Action
Patients
Community
Community
10
Example of standardisation, one room in one
health centre, Mitchells Plain, South Africa
11

• Standards What are the relation between
  technical
• Standards like database tools network protocols
  and
• data standards ???
• HISP in South Africa has basically focused on
  data standards
• Data elements indicators
• The District Health Information Software DHIS
• The perspective on this may be that
• Data standards are relatively easy to change
  (although inertia ..)
• And that the DHIS illustrates a very flexible way
  to implement these standards
• This is linked to layering

12
- Standardisation II in the IS field
basically about technical integration - Need to
be extended to all layers?
Social Social /Technical sphere
Technical sphere
The 7 layers ISO model for communication
standards Logical / application layers (e.g.
communication between databases) protocols
for communication (which bits first
etc.) Compatible plugs Physical layer,
telephone lines
Need to expand logical layer Shared
understanding (of purpose, needs, etc.) Shared
terms (to express this) Language (English,
Esperanto ) Shared categories (formalised for
communication) Formal languages (application
layer)
Systems
13
Distinguish between layers An airport (physical
layer) is more cast in stone than a verbal
agreement (data or higher layer)
The 7 layer OSI model for technical layers data
layers organisational layers
Socially /institutionally embedded standards
Organisational /social layer
More Flexible?
Data layer (content, meaning, logic)
Data layer more easy to change a data
definition than the technical infrastructure
Application layer (databases, GIS, ..)
E-mail / web layer (http, ftp)
Communication layer (protocols, TCP/IP)
Hard coded standards
Physical layer (network)
Less Flexible?
Physical standards
14
1. Use (organisational) dimension
Organisational Practises Information Needs
use

• Three layers / dimensions of
• information systems management
• Use provide users withneeded information
  inappropriate formats
• Data establish and maintainessential
  data/indicator sets standard data procedures
• Tools establish maintain appropriate tools
  infrastructurefor data collection, flows use

Use of information in organisational context
Interacting with the information system
Accessing /producing information
2. Data layer
Data indicators for use
Definitions of data and indicators. Data about
data meta data, e.g. data dictionary
Standard indicator data sets Data elements
3. Technical/tool layer (NOTE uneven/mixed
infrastructure in Developing Countries)
Applied to the mixed/informal reality of
developing countries (paper/computer)

• Tools
• Infrastructure
• for
• Data
• - collection,
• - flows
• - management
• analysis
• presentation
• - reporting

Layered model for software /systems development
(developed infrastructure!)
Application layer databases etc.
Paper based registers, archives, tools AND
computer spreadsheets, databases
E-mail / web layer systems for carrying
/presenting data (e.g. http, ftp)

• Communication data flows
• Paper forms, files, CDs, flash discs, web
• Ad hoc transport, courier, postal services,
• Internet, sms

Communication layer protocols and network
standards (e.g. TCP/IP)
Physical layer telephone lines, fiber etc.
General infrastructure for data flows From roads
to Internet mobile network
15
International
Standard Data, Datasets Indicators For each
level
National
International level
Province
National level federal
District
Sub-national level province, region
Community
District level woreda, local government
Patient / Singular unit
Health Facility community level clinic,
health centre, hospital
Patient level or singular, local unit for not
patient related data
Hierarchy of standards data layer
16
Uneven development of Information infrastructure
in developing countries Challenge because the
technical solutions must be different Standardisa
tion must focus on data layer!!! The technical
layers must be solved through ad-hoc gateways and
flexible solutions (from web to donkeys!)
Addis Ababa
Oromia
Benishangul
Levels
Region
Zone
Worreda/ Sub-city
Health centre
Patient / Primary registration
Computer
Gateway
Paper
17
South Africa Standardisation at the data level
shared essential minimum approach But highly
supported by DHIS, standardisation at application
level
Higher levels
Health programs
Family Planning
Mother Child
Database Info. office
A
B

• A) Pre-apartheid vertical and fragmented
• structure in Atlantis (simplified)
• - no local use analysis of data
• 172 data collecting forms in 3 gov. structures
• multiple health programs (immunisation, TB,
  etc.)
• Data overlaps and gaps and inconsistencies
• Data handling takes up to 40 of nurses time!

• B) Ideal integrated district model
• South African HISP approach was to
• create a common minimum data set,
• one sheet of paper covering key
• Issues, addressing key indicators
• - This seemingly logical approach has
• Worked in NONE of the other
• HISP countries! (why?)

18

• South Africa (1)
• Post apartheid South Africa extremely fragmented
  health services (13 Department of health, all
  with different reporting structures and
  standards)
• Inequity disadvantaged groups, blacks, former
  homelands very poor health status and health
  services
• 1994- 95 quest for equity (everybody wanted
  changes) and an enabling environment for change
• Federal structure each province could innovate
  and go their own ways
• HISP started in one then two provinces
• Minimum data set approach focus on Use /
  indicators
• Combined with
• A very flexible database application approach
• Combined prototyping of datarototyping of HISP
  in South Africa has basically focused on data
  standards

19

• South Africa (2)
• Implementation in two provinces showed very quick
  results
• Proved effective approach
• Success in two provinces convinced national and
  other provinces
• Important that programs and provinces did not
  feel threatened
• could continue more or less as before, as long
  as they adhered to the standards of the level
  above
• NOTE the two provinces implemented different
  standards!
• The success of the approach, the way all actors
  understood how they could take part in the
  process, maybe more than the pure results, was
  the important factor
• In other countries standardisation following the
  blueprint of South Africa has not succeeded Less
  of a federal structure, less ready for change,
  more data need to be included, more overall
  system approach

20
Data flows reporting from facility to central
level in Botswana converging at district level
- Fragmentation at central level
IDSR Notifiable Diseases
EPI
Home Based Care
ARV
Health Statistics
PMTCT
STD
Nutrition
Nutrition
Family Planning
MCH
HIV/AIDS
TB
School Health
Mental Health
And more
District - DHT
Facility 1
Facility 2
Facility n
Facility 3
21
IDSR Notifiable Diseases
EPI
Home Based Care
ARV
Health Statistics
PMTCT
STD
Nutrition
Nutrition
Family Planning
MCH
HIV/AIDS
TB
School Health
National HMIS Stat. unit
Mental Health
And more
District HMIS
Facility 1
Facility 2
Facility n
Facility 3
22
PROBLEM From district to central - From
integration to fragmentation
Safe motherhood
Zone (?)
District North A
Malaria
Statistics unit
Nungw
HIV/AIDS
Drugs
Gomani
Kijini
DHMT
Jongowe
TB
EPI
Matemwe
Kivunge
Mkokoti
Pwani mchangani
Nutrition
Chani
Human resources
STI
ETC.
23
Integrated HMIS - Proposed solution
Other stake holders
National Statistics
Other directors
District North A
Minister Of Health
Dir. Planning
Nungw
Gomani
Kijini
District HMIS
Central HMIS
Jongowe
Zonal HMIS
Matemwe
Kivunge
Mkokoti
Safe motherhood
Statistics
Pwani mchangani
Chani
EPI
Malaria
Nutrition
TB
Human resources
Drugs
HIV/AIDS
STI