The challenge of heterogeneity in HMIS: Addressing the challenge

1
The challenge of heterogeneity in HMIS
Addressing the challenge

• Sundeep Sahay
• NHSRC

2
Overview

• Elaborating on heterogeneity and HMIS
• Ongoing approach to address it in our national
  context

3
The nature of HMIS

• Airplanes do not fly airlines do....

4
Implying.....

• Technology is only one component (often glorified
  over others!!) of a ----
• complex, heterogeneous. socio-technical and
  inter-connected network
• There is always history and legacy you cannot
  ever design from a clean slate
• Changing one part of the network always effects
  some other part

5
Unpacking heterogeneity

• The facets of heterogeneity
• multiplicity of informational needs.
• hierarchies of standards.
• multiplicity of institutions and organizing
  forms.
• variety of work practices surrounding the HMIS.
• multiplicity of technologies and legacy systems
• complexities of the socio-political- historical
  conditions

6
Information ongoing tensions

• Variety of data types
• routine
• survey
• infrastructure
• staff
• epidemiological etc
• District versus program needs
• Varying periodicities of collection and reporting
• Donor versus district needs
• Dashboard versus systemic needs

7
Variety of work practices

• Multiplicity of accountabilities
• Social construction of deadlines
• Variations in resources
• Varying administrative and program needs
• Private sector, little no incentive to report,
  even a disincentive
• Geographical spread of work

8
Different historical trajectories

• Historically entrenched actors interests
• HMIS high political visibility rapidly changing
  political scene. NRHM acknowledges priority of
  HMIS
• Politics of diseases
• Big money large cake especially for esoteric
  technology projects (eg EPR, telemedicine, voice
  recognition)
• Unequal distribution of infrastructure

9
Some examples
10
Information flows parallel and multi-level
11
Varieties of technologies (and paper)
12
Multiplicity of interventions (within HIV/AIDS)
13
Work practices surrounding ART System
14
Heterogeneity of Actors (HIV in Ethiopia)

• International donors
• CDC
• John Hopkins University
• Colombia University
• John Snow International
• Government
• FMOH
• Program managers
• HAPCO National agency responsible for HIV
• Other NGOs
• University researchers - AAU

15
Approaches to address heterogeneity in HMIS

• Typical approaches
• HMIS is inherently messy so eliminate
  heterogeneity
• make all encompassing systems where one shoe
  fits all
• Using latest technologies (example web systems)
  will kill the mess
• But, heterogeneity
• is part of life, cant eliminate
• comes with both opportunities and opportunities,
  should also be celerbrated
• also, cant live with the mess
• need to find way to make the mess work a
  pragmatic balance

16
What are we doing?

• Elaborate on ongoing work at the MoH

17
Briefly Existing situation

• Data related
• unreasonable number of data elements
• high of blank or zero values
• duplications and gaps systemic ambiguities
• indicator to data mismatch
• lack of uniformity standards in naming
  conventions
• Systems related
• disproportionate attention to hardware/software
• undue burden of reporting on field staff
• weak support for supervision and feedback
• inadequate quality assurance mechanisms
• fragmentation and verticality of systems

18
Unreasonable numbers of data elements

• For example
• Gujarat District 1128
• Kerala SubCentre 1667
• Jharkhand
• SC 165
• PHCs 458 (SC/ST break ups not provided)

19
High proportion of 0/blank values
20
Duplications and gaps systemic design
ambiguities

• Data duplication - a minefield
• Field workers report data both on services they
  provide AND institutional services
• RIMS and IDSP repeat same data as form 6
• Fragmentation by programs
• parallel collection system for programs
• RIMS, IDSP
• Same data on different forms
• e.g. Form 6, RIMS, IDSP
• Gaps in data reports
• e.g. BEOC, Quality of Care, HIV, Laboratories

21
Data element-indicator mismatch

• SC/ST/Others tripled previous data elements,
• yet of SC coverage not compared with other at
  State or National Levels
• Male/ female adds 75 data elements for
  immunization,
• yet sex difference in immunization coverage not
  reflected in plans, eg PIPs
• Gujarat is exception in this regard

22
Excessive numbers of elements in numerators

• Some examples
• of deliveries attended by SBA 40
• Prop. of institutional deliveries 36
• Met need of EOC 37
• Couple year protection rate 245
• Serious implications on the correctness of
  indicators calculated
• Estimating denominators is non trivial challenge

23
Lack of uniformity in naming of data elements

• Different states have different names for same
  activity
• Number of pregnant women given prophylaxis for
  Anemia (Kerala)
• Number of pregnant women given 100 IFA tablets
• lost (and gained) in translation
• Activity described rather than uniform and
  concise variable names

24
Disproportionate attention to hardware/software

• State HMIS efforts typically spend
• 75-90 on hardware/software
• 10-25 on capacity building and implementation
  support
• The proportion should be the reverse
• Use of proprietary software locks the Health
  Department to the vendor, and with it expensive
  efforts to incorporate any change
• Some expensive efforts to use hand held devices
• Use of Free Software needs to be further
  explored. Currently, positive results from
  Gujarat and Kerala

25
Undue burden on field staff

• Excessive forms and data
• Multiple primary registers (more than 20)
• Multiplicity and redundancy same data in
  different forms
• ANM expected to report on items for which not
  equipped
• eg Diphtheria, Child TB, Hepatitis
• What takes a backseat is
• quality of data
• use of HMIS for local action should be the
  raison detre of a HMIS

26
Approach to dealing with this heterogeneity

• Establishing coherent design principles
• overall design
• data gathering
• indicators

27
Overall design

• HMIS should be action-led, not data-driven
• Action led means
• careful choice and definition of indicators
• emphasis on processes around their analysis,
  interpretation and action
• Focus to shift from FORMATS and DATA ELEMENTS to
  INDICATORS and their use for action
• Each data element collected should contribute to
  at least ONE indicator, preferably MORE
• A concrete mechanism to reduce field staff
  workload

28
Data gathering

• Service providers report only data from facility
  where service provided
• Increase primary record reviews, local surveys
• SC/ST, Age, Gender, BPL etc Local surveys,
• regular cross checking routine data for improved
  quality and strengthening supervision
• increase downward flows of information

29
Defining hierarchy of indicators
30
Implications

• Core National level indicators
• obligatory for levels below to provide data for
  core indicators
• successive levels can add or delete indicators as
  long as they dont affect the core

31
Routine HMISSupporting a variety of functions
Input
32
Using multiple data sources
33
Results of applying these principles

• Dramatic reductions of data elements
• Sub Centre From about 1500 to 55
• PHC From about 500 to 140
• CHC From about 600 to 150
• Each data element linked to at least ONE
  indicator
• Some of the redundancies are in the process of
  being removed. eg
• Sub Center and RIMS
• Sub Center and IDSP
• Sub Center and TB

34
Design framework for technical integration data
warehouse approach

• Unified System for
• Reporting
• Dashboard
• Data Validation
• GIS

Malaria
RCH
HIV/AIDS
RNTCP
Others
NLEP
Various Systems in their individual
formants/database
35
Design framework for institutional integration
and information flows
Malaria
TB
Key Functions of HMIS unit 1. One point for
compilation and entry of all routine data 2.
Generation of all routine reports and indicator
based analysis 3. Transmission of data and
reports vertically to HMIS unit higher level 4.
Transmission of reports horizontally to different
centers (TB, Imm etc) 5. Serve as HMIS training
unit for that level
RCH
HMIS unit Dataware house
NLEP
IDSP
Blindness control
Others
36
Moving from design to practice

• A non-trivial challenge
• Our approach in gradual steps
• field consultations in various states
• on adequacy of the data elements
• understanding compatibility with primary
  recording registers
• ease of use
• national level consultations with ME division
  and divisional heads on the adequacy of the
  indicators
• Creation of supporting documents
• data dictionary for data elements
• data dictionary for indicators
• guidelines for use
• Designing of reporting formats
• Developing institutional and political buy-in at
  various levels (MoH, States, Development Partners
  etc)

37
Various next steps

• Establishing principles for HMIS design action
  first step
• When agreed upon, process gradually needs to
  extended to other NRHM programs (TB, Malaria,
  NLEP etc)
• Initiate implementation processes supported with
  capacity building, documentation
• Registers need to be rationalized at field level
• Integration of register data with the sub centre
  data
• Exploring the use of hand held devices (eg mobile
  phones, digital pads) for supporting primary
  registration of data and their integration with
  sub centre data
• Developing strategies for scaling up and the
  sustainability of HMIS related processes
  technical and institutional