improving the quality and use
TRANSCRIPT
-
8/10/2019 Improving the Quality and Use
1/20
Strengthening health systems
in Asia and the Pacifc through
beer evidence and pracce
For the PDF version of
this paper and other
related documents, visit
www.uq.edu.au/hishub
Theme: Building health informaon systems
Working Paper Series Number 5 November 2009 WORKING PAPER
Improving the quality and useof health informaon systems:essenal strategic issues
Health Informaons System Knowledge Hub
Instute for Health Metrics and Evaluaon
-
8/10/2019 Improving the Quality and Use
2/20
-
8/10/2019 Improving the Quality and Use
3/20
1Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
Contents
Acronyms and abbreviaons ................................................................................................................... 2
Introducon............................................................................................................................................. 3
Recent developments in strengthening health informaon systems ..................................................... 4
The health informaon system as a stascal tool ................................................................................. 7
Health data sources ................................................................................................................................. 9
Conclusion .............................................................................................................................................. 14
References .............................................................................................................................................. 15
-
8/10/2019 Improving the Quality and Use
4/20
2 Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowledg
eHub
Acronyms and abbreviaons
DHS demographic and health surveysHMN Health Metrics Network
ICD-10 Internaonal Stascal Classicaon of Diseases and Related Health Problems,10th revision
MDG Millennium Development Goal
RHINO Roune Health Informaon Systems
WHO World Health Organizaon
-
8/10/2019 Improving the Quality and Use
5/20
3Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
The paper is intended to provide a succinct overview
of issues that countries and the donor community
might wish to consider when developing strategies
and pracces to improve the quality and use of
health informaon.
This paper sets out some of the crical issues that
countries and donors should consider when invesng in
the development of health informaon systems. These
range from incenves and pracces to improve the
quality and, especially, the use of health informaon by
those in policy who have the greatest need for reliable,
mely and relevant health informaon for planning, to
strategies to create a culture of informaon demand
and use. Experiences from countries such as Mexico and
Brazil have been analysed to draw aenon to pracces
that might be protably adopted elsewhere, parcularly
in the AsiaPacic region. The paper includes a cursory
overview of internaonal iniaves in health informaon
strengthening. It also reviews the crical role of various
health data sources in supporng indicator development
for the monitoring and evaluaon of health status and
health program eecveness. The paper is intended
to provide a succinct overview of issues that countries
and the donor community might wish to consider when
developing strategies and pracces to improve the
quality and use of health informaon.
Introducon
-
8/10/2019 Improving the Quality and Use
6/20
4 Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowledg
eHub
to produce high quality primary data and to make this
data available for measurement purposes. Data are a
rst step, but quality data is the real aim. The glossary
for evidence based public health (Rychetnik et al. 2004)
denes data quality as, the degree to which data items
are accurate, complete, relevant, mely, suciently
detailed, appropriately represented (e.g. consistently
coded using a clinical coding system) and retain sucient
contextual informaon to support decision making.
These characteriscs that dene data quality should be
taken as principles for any health metrics eort, and at
the same me, serve as a challenge for HIS.
Incenvising quality and use
For a typical health informaon system in a developing
country, it is not easy to achieve data quality. Frequently,
health data in developing countries are incomplete
they either miss a poron of the populaon or do
not cover all relevant aspects of health. This is oen
through no fault of their own; they simply do not have
the resources needed to achieve a comprehensive
system instantaneously, but they can denitely work to
improve what they have. There is also somemes a lack
of support from the supply perspecve for improvingdata quality. There are few incenves to correct the
crude data gathered for the health informaon system
at a naonal or district level. This generates a perverse
cycle in which decision-makers reacng to the quality
problems in the data exclude those data from their
decision-making. In turn, providers of data choose not
to invest in improvements because nobody is consuming
their products to begin with (Ash et al 2004).
To break this cycle, it is necessary to create incenves,
both to use beer informaon at the local level and
for providers to deliver high-quality, mely data. Agood starng point would be for the central health
informaon system to require data to be disseminated
on a clear schedule. However, this would not be enough
to ensure that local providers are using the informaon.
In other words, ensuring high-quality data is a necessary
precondion for geng that informaon used by
decision-makers and praconers, but it is in itself,
not enough.
For a typical health informaon system in a
developing country, it is not easy to achieve data
quality. Ensuring high-quality data is a necessary
precondion for geng that informaon used by
decision-makers and praconers, but it is, in itself,
not enough.
Health metrics and health informaon
systems
Murray and Frenk put forth the term metrics, to refer
both to the science of measurement and to a specic set
of instruments and indicators that provide the empirical
basis to understand a parcular object of enquiry
and acon (Murray & Frenk 2008). According to this
denion, health metrics is related to measurement of,
and to the instruments used to measure the health of
the populaon. Measurement refers to esmang the
magnitude of some aribute of an object (health status)
and usually includes measuring tools, such as stascal
models, which are calibrated to compare the object
against an established standard.
It is important to think of health as the product of many
intertwined factors. A systemic approach to gathering
and assessing health informaon is therefore paramount
to good decision-making. This perspecve and approach
is quite common in medicine for clinical or individual
purposes. However when the problem concerns the
organizaon, community or general populaon, the
approach is oen fragmented. In the face of this
challenge, technical soluons may be perceived as the
cure-all, when in fact a more systemac diagnosis
and treatment is needed. The systemic approach forpublic health problems should consider simultaneously
human resources, the organizaonal environment and
processes, and technology - and how to align them. In
other words, besides the technical transformaon of
the health system, an important cultural change in its
organizaon needs to happen. It cannot burden the
informaon providers to a degree that it is resisted, or
be so cumbersome for an informaon consumer that its
results are ignored.
The work to be done on the health informaon
system side is inherently connected with the idea of
measurement. Any HIS should have as two of its goals
Recent developments in strengtheninghealth informaon systems
-
8/10/2019 Improving the Quality and Use
7/20
5Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
good-quality data. There is also a scarcity of incenves to
use the informaon collected through health informaon
systems for decision-making and poor use of evidence
from any source for decision-making.
There is a very useful tool developed by the Roune
Health Informaon Systems (RHINO) to idenfy which
aributes of health informaon systems directly
contribute to developing poor-quality products (Aqil
et al 2009). The instrument used the Publishing
Requirements for Industry Standards Metadata (PRISM)
framework, which organises the determinants of the
health informaon system performance into three
blocks: technical, organisaonal and behavioural.The instrument was designed to be applied at a local
level, and the results obtained, unl now, have been
used more for local consumpon. Nevertheless, it is
possible, and important, to generalise some lessons
aer its applicaon in South Africa, Tanzania, Pakistan
and Mexico. In relaon to the organisaonal and
behavioural determinants, two main problems have
emerged as common in all countries: the lack of a culture
of informaon, and the immense need for in-country
capacity building.
Conversely, there are some experiences thatdemonstrate that a posive and supporve culture
can be constructed around the producon and use
of informaon using the products of the current
health informaon system. For example, since 2001,
the Ministry of Health in Mexico has been publishing
an annual accountability report called Salud-Mexico
(Secretaria de Salud 2002). The goal of this report is to
document the states performance benchmarking system.
The model adopted to present the informaon was
based on selecng mainly health outcome indicators.
The report uses the past year (or somemes the past two
years) as reference points to measure improvements in
the health system.
The Mexican Government created incenves for
informaon providers by mandang that the report
be released publicly in a cizens forum, which brings
together important federal and state decision-makers,
civil society leaders, academics and the media. The
government presets a date for the release of the report
to provide a clear deadlinethe beginning of the second
quarter of the year. Taking this approach created strong
incenves for informaon providers to complete data
collecon, data processing and data integraon withina short period. Because it was a naonal accountability
Creang a culture of informaonand building capacity
Building capacity in a country, like fostering a culture
of knowledge, requires a good understanding of the
operaonal environment. In many countries, building
capacity requires reorganising past methods for
informaon collecon. For example, for some naonal
surveys, if the sample process is not standardised, the
results are not comparable. This disconnuity allows
health informaon system managers and other decision-
makers to operate in dierent dimensions; they may
only focus on immediate results or specic areas andnot examine naonwide trends. A lack of harmonisaon
between naonal surveys also results in a weaker
informaon system because managers and decision-
makers have no uniform way to hold their systems
accountable at a larger scale. Also, the capabilies of
health informaon system workers are more dicult
to understand without a uniform test to apply. With
more rigorous programs to build worker capabilies
and program capacity, health informaon systems could
become stronger both within and across countries.
Another reason that capacity is somemes weak is that
health informaon systems do not span across a Ministryof Health or equivalent. Instead, there are somemes
only small units of informaon and informacs within
the health programs, but no organised systems to bring
them together. In order to scale up the abundance
of data across units, branches and departments, it is
necessary to highlight the importance of having a logical
and transparent structure; the importance of integraon
to serve all users; and the importance of keeping
autonomy regarding any kind of informaon (avoiding
conicts of interest) (Krikeberg 2007). If these principles
were applied to the informaon gathering systems by
ministries of health in a comprehensive and systemacmanner, the quality, use and usefulness of informaon
would increase markedly.
The lack of a culture of informaon is widespread. This
concept can be dened through several selected domains
by aempng to understand the degree to which a
country: uses numbers to describe problems and their
soluons; aempts to understand a problem through
the collecon of data and informaon; establishes a
connuous quality learning process; and empowers
people through imparng informaon and knowledge.
In summary, in developing countries, there is oen anabsence of a common commitment to, and support for,
-
8/10/2019 Improving the Quality and Use
8/20
-
8/10/2019 Improving the Quality and Use
9/20
7Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
management demonstrates the various capacies in
which health informaon systems have been used.
The HMN developed a parcular health informaon
system assessment framework to help countries assess
how best to improve their systems. This framework
adapts key components arculated by the General
Theory of Systemsinputs, process, outputs and
boundaries. As such, it does not prescribe precise,
step-by-step requirements, but rather what the general
components are and why they are needed. The HMN
framework proposed a new structure in which health
informaon system resources, not data, are the input.
Specically, the structure denes input as the legislave,regulatory and planning framework required to ensure
fully funconing health informaon systems, and the
resources to ensure that each system is funconal
(personnel, nancing, logiscs support, and informaon
technology and communicaon).
The framework idenes three key components in
the process of strengthening the health informaon
system for a parcular country. The components are
idenfying indicators, data sources and methods for data
management. According to this approach, collecng,
storing, processing, compiling and analysing the data areall integral parts of well-funconing health informaon
system. The outputs include two key components:
informaon products, and disseminaon and use (ie the
transformaon of data into informaon that will become
the basis for evidence, and how to make the products of
the system available and accessible to decision-makers).
The important connecon between this framework and
how health informaon systems can be used by decision-
makers has been presented by Lozano et al (2007)
in the context of assessing eecve coverage for key
intervenons. Three specic lessons were arculated,which could be generalised to other types of health
informaon:
1. Naonal or local household health surveys are
good data sources to measure eecve coverage;
however, they must be supplemented by other
techniques to measure some aspects (eg quality).
2. Roune registries can also be very good data
sources; however, registries should be at the
individual level and have a high degree of accuracy.
In the coverage example, follow-up of people with
Health informaon systems are dened in many
dierent ways, depending on who is using them versus
who is implemenng them. Although these denions
vary, health informaon systems more generally
are the nexus of informaon, technology and the
accompanying processes to provide strategic access to
informaon for decision-makers. Health informaon
systems are comprised of resources, mechanisms
and methods that facilitate the acquision, storage,
retrieval and use of data in health and health decision-
making. The role of a health informaon system is to
determine what informaon needs to be collected and
tracked; to establish mechanisms for collecng the
informaon; to build and sustain an ongoing process of
adding value to the data collected; to ensure that the
data are understood and used; and to substanate the
need for data collecon so that funding is maintained.
Health informaon systems can have mulple
aributes. They can be: paent centred or public health
oriented; subject based (paents, doctors, etc) or task
based (hospital discharges registries); paper based or
computer based.
In any of its contexts, a health informaon system
should act as value chain with dierent components
that transform the original facts gathered into
knowledge that can be applied by policy-makers to
improve populaon health. While there are diering
opinions in the literature about the scope of health
informaon systems, there is general agreement that
they are complex, dynamic, context based and of great
social importance (Lenz et al 2002, Detmer 2003).
To capture this informaon in a more standardised
and comparable way, the use of health metrics has
been increasing. Most recently, exibility of health
informaon systems has been highlighted as an
important component. This exibility includes beingable to receive and store data from many dierent
sources and from mulple dimensionsfrom individual
paent informaon to populaon-level me trends of
morbidity; from storing only alphanumeric informaon
to storing medical diagnosc images. Other recent
changes include the heightened consideraon of health
informaon system users. This is evidenced by the
inclusion of paents and health consumers and through
diversifying the use of data beyond paent care and
administrave purposes. The shi from focusing mainly
on technical health informaon system problems to
those of change management and strategic informaon
The Health informaon system as astascal tool
-
8/10/2019 Improving the Quality and Use
10/20
8 Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowledg
eHub
Network for Health Informaon (with the acronym RIPSA
in Portuguese). The consensus process was arduous and
lasted about six years. However, the product delivered
was a manual that included detailed analyses about the
quality of each database, key indicators to track, and
common denions and ways to measure them. The
agreement included a naonal mandate about the use
of crude and corrected stascs when reports are used
to describe subnaonal levels. It included descripons of
the data source to be used and the methods to correct
completeness problems. RIPSA connues to operate
today as a valuable process to ensure high-quality,
standardised indicators (RIPSA 2002).
Geng naonal agreements about the inclusion of
corrected stascs as an input for decision-makers is
crucial. In Brazil, it was possible because there was a
crical mass of experts that were able to employ these
stascal methods and to engage in nuanced and
informed arguments with the naonal group charged
with making informaon available and useful. They were
even able to strategically think about useful indicators for
measuring progress with MDG such as infant mortality
and maternal mortality. This opened the door to bridging
the gap between informaon and policy.
AbouZahr et al (2007) proposed a schema to smooth
the pathway from informaon to policy. Their
recommendaons for good pracce in the informaon
producon process should be considered seriously
because they apply to all health systems and to both
developing and developed countries. According to
the WHO framework for Health Systems Performance
Assessment (Murray and Frenk 2000, WHO 2000),
health informaon is one of the key components of
the stewardship funcon of health systems. To support
key acvies such as priority seng and performance
assessment, high-quality informaon is a fundamental
ingredient. Today, health informaon is becoming more
important and is idened as a fundamental building
block of the health system (WHO 2007). It is as important
as service delivery, the health workforce or nancing.
AbouZahr et al (2007) also raise the importance of the
tension that internaonal agencies usually generate for
countries when the local and internaonal indicators
do not match. In a country with a decentralised health
system, a similar problem also arises, though on a
smaller scale, between the federal/naonal and the
regional or local authories. Another obstacle can be amisunderstanding about ocial data. For many people,
chronic condions such as high blood pressure and
diabetes, and other acute condions, is necessary.
3. Fragmentaon of the health system generates a
fragmented health informaon system. Eecve
coverage is a metric that is useful in combining
informaon from both the public and private sectors
of the health system.
The use of informaon with sucient quality for
comparability purposes requires establishing a broad
(naonal or internaonal) and widely accepted
mechanism of standardisaon. Somemes, there is
tension amongst the involved groups that works against
such acceptance of standardisaon. As informaon
becomes more global, this problem will likely
receive more aenon. TheInternaonal Stascal
Classicaon of Diseases and Related Health Problems,
10th revision (ICD-10), is a good example of a stascal
tool that creates constant tension between universal
standardisaon and local circumstances. However, this
tension cannot be resolved by sheer force because the
problem is persistent. For example, whenever there
is an update to the ICD, to ensure the highest degree
of comparability between countries each individualcountry must comprehensively update their systems and
retrain their health workforce to code deaths using the
updates. The ICD creates updates because it wants to
incorporate latest advances in medical knowledge, but
the transional cost to individual countries can be high.
Health informaon systems must have standards, but,
at the same me, countries should have the exibility
to adapt to changes in their own health systems. The
need for both standardisaon and exibility must be
balanced in any health informaon system. In many
instances, these opposing aributes have helped
drive reform discussions. For example, in South Africa,
standardisaon of health data was a major element in
the process of changing the health informaon system.
In Brazil in 1995, an ad hoc group was formed to develop
and agree on a way to design databases at dierent
instuons to help store informaon to calculate a
core set of agreed indicators. Representaves from the
main areas of the Ministry of Health, key parcipants
from the informaon instuons of the country, and
representaves from universies and academia all played
a role. The Pan American Health Organizaon (PAHO)
was also an important player in this iniave (PAHO2001). The project became known as the Interagency
-
8/10/2019 Improving the Quality and Use
11/20
9Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
Health data sources
At a more general level, health stascs can be classied
under two broad categories: primary microdata, and
combined (aggregated) secondary macro-datasets.
Several typologies exist that capture specic types of
health data within these main categories.
Primary microdata
This is the category of major importance for generang
health stascs and is the foundaon of most other
datasets.
Vital registraon
Vital registraon data collects birth and death
informaon and, given their importance, provides some
of the most advanced and standardised types of health
data across countries. The goal of any vital registraon
system is to accurately record allbirths and deaths in the
populaon as a whole. We can classify vital registraon
sources into four dierent types:
Complete vital registraon systems, including cause-
of-death cercaon and coding according to the
WHO ICD-10. For this system, the death cercate can
provide crical data, parcularly if the informaon
is digised along with the ICD-10 codes. These data
become even more useful if they include a naonal ID
number on the death cercate that allows matching
to other data collected in household surveys,
censuses or health service registries.
Incomplete vital registraon systems that collect
cause-of-death informaon according to the ICD-10,
but are incomplete and/or do not have reliable death
cercaon or naonal ID systems in place.
Sample registraon systems, where vital registraon
has been implemented in a sample of localies
that are intended to be representave of the enre
country. These systems have been successfully
introduced in very populous countries such as India
and China.
Demographic surveillance sites (DSS) are used in
many low-income countries to monitor vital events
in specic communies. DSS were oen inially
developed as research sites and have been used
for large-scale clinical trials. However, a number of
them have persisted and diversied the informaonthey collect. The largest network of such sites is the
ocial data by denion are those gures produced
by government agencies, regardless of the quality of
the data. If the informaon comes from the Oce of
Stascs of the Ministry of Health, it is automacally
considered to be ocial and ready to use. However,
informaon released by a government agency must sll
be scrunised for quality.
Despite this tension, from a country perspecve,
it is important that a health informaon system
responds primarily to country needs and secondarily
to internaonal needs, such as those of the WHO.
Even aer these priories have been sorted out, data
ownership is not a minor issue, and is likely to be one ofthe most common sources of misunderstandings and a
signicant obstacle to successful collaboraon between
naonal and internaonal agencies. When there is no
clear ownership of informaon, there is likely to be
tension. The problems arise when the calculaons from
internaonal agencies produce esmates for the same
indicator that dier from those of naonal agencies.
Somemes, it is dicult methodologically to explain
why the two might be dierent. However, even if there
is good reason for the two to dier methodologically,
it is even more dicult to explain polically, and to a
nontechnical audience, why a predicted value from an
internaonal agency is much higher or lower than the
ocial data of the country. The negave consequences
of these problems generate two dierent reacons.
Naonal stakeholders, believing that the internaonal
organisaons are not acng in their best interest, will
not share their data. Subsequently, the internaonal
organisaons end up using incomplete databases, which
forces them to model more esmates with limited data
and simply perpetuates the original problem.
-
8/10/2019 Improving the Quality and Use
12/20
10 Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowledg
eHub
of health surveys in the United States. The objecve
of the BRFSS is to collect uniform, state-specic data
on health risk behaviours, clinical prevenve health
pracces and health care access that are associated
with the leading causes of morbidity and mortality.
Data are collected from a representave sample in each
state, and the sampling is designed to provide naonal
esmates when data from all states data are combined
(Mokdad et al 2003).
Fourth, disease-specic or condion-specic
standardised surveys, oen sponsored by WHO to
examine a parcular health problem, have been carried
out in many countries of the world. These surveys, whichare relavely standardised due to WHO sponsorship,
cover malnutrion, adolescent risk factors, oral health,
etc. A good example is WHOs Global Database on Child
Growth and Malnutrion, which comprises informaon
collected from populaon-based naonal and
subnaonal surveys that follow a standard procedure
to obtain comparable results from about 155 countries.
In Australia, the Naonal Survey of Mental Health and
Wellbeing (SMHWB) is designed to collect informaon
on the mental health and wellbeing of the Australian
populaon. The objecves of the survey are to provide
informaon, about Australians aged 18 years or more,
on the prevalence of selected major mental disorders,
the level of disability associated with these disorders,
and the health services used and the help needed
as a consequence of mental health problems. The
informaon is collected by personal interview from usual
residents from approximately 15 000 private households.
Fih, verbal autopsy is a highly specialised survey to
esmate cause-of-death paerns in populaons that
lack vital registraon. The survey focuses on collecng
informaon from families of the deceased about signs
and symptoms before the death. It can also be used by
trained health professionals to idenfy causes of death
in environments where most individuals die at home
or without any contact with medical establishments.
The verbal autopsy method has been developed to ask
relaves of the deceased a series of symptom-based
quesons about the events leading up to death, as well
as broader sociodemographic and risk factor informaon
that might yield clues as to the cause of death of the
deceased. Based on this informaon, a cause of death
can be assigned by a clinician. Although potenally very
useful, there is less standardisaon in verbal autopsy
instruments that have been used. However, there
have been eorts to improve their comparability. Also,
Internaonal Network of Field Sites with Connuous
Demographic Evaluaon of Populaons and Their Health
in Developing Countries (INDEPTH). Because they are
not designed to be naonally representave, these
sites are less useful to monitor naonal health levels.
Household interview surveys
Household interview surveys have followed standardised
instruments and protocols, beginning with the World
Ferlity Survey in the 1970s. It is useful to think of them
under several dierent headings according to their
main purpose.
First, there are a set of fairly standardised mulcountry
health interview surveys. These are not standardised
with one another, but each survey program is
implemented in approximately the same form in mulple
countries. These include the demographic and health
surveys (DHS), the World Health Surveys, the mulple
indicator cluster surveys (MICS), the Pan Arab Project
for Family Health (PAPFAM), the Pan Arab Project for
Child Development (PAPCHILD) and the United States
Centers for Disease Control and Prevenon (CDC)
Reproducve Health Surveys. These surveys provide
valuable informaon across and within countries,parcularly because each survey is fairly standard across
the countries in which it has been implemented. For
example, each DHS is fairly comparable to another,
regardless of the country or the year in which it was
implemented. The surveys are also useful because of
the amount of specic health details that can be elicited
from these interviews.
Second, there are standardised mulcountry surveys, the
primary purpose of which is collecng socioeconomic
data, but which also have health modules. Examples
include the Living Standards Measurement Surveys(LSMS) and the United Naons Household Surveys.
The LSMS includes detailed health quesons that are
helpful in measuring the ulisaon of health care and
health expenditures.
Third, naonal household health interview surveys are
somemes ulised by countries to track health issues
and levels within their own countries. These surveys
allow for standardisaon over me and are specically
tailored to a countrys health prole. In the United
States , the Naonal Health Interview Survey has been
conducted for over 30 years. The Behavioural Risk FactorSurveillance System (BRFSS) is a state-based system
-
8/10/2019 Improving the Quality and Use
13/20
-
8/10/2019 Improving the Quality and Use
14/20
12 Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowledg
eHub
Budgets and expenditure reports
Expenditure and budgetary data provide valuable
informaon on nancial resources for health. These data
come from naonal budget documents, expenditure
reviews and audit reports. Most oen, this informaon
is available at the summary level. As part of the
management of health services, budget and expenditure
data are frequently provided by nancial management
informaon systems, which are somemes maintained
for the government as a whole, rather than just for the
health system. There is less informaon being collected
on individual expenditures on health, but these data are
certainly desirable to obtain when possible.
For policy development and strategic planning, nancial
data are oen compiled using the methodology for
naonal health accounts (NHA) (OECD 2000, World
Bank et al 2002, Poullier et al 2003). This system
provides informaon on the amount of nancial
resources available for health and their ows across
the health system. The breakdown of data into private
and public sector categories is an important aspect in
this regard. In addion, the disaggregaon of nancial
informaon by major disease or health program area is
possible. At subnaonal levels, budgetary informaonlinked to health system funcons and, in parcular,
health intervenons, is a minimum requirement for
performance budgeng.
Epidemiological observaonal studies
Epidemiological observaon studies follow a cohort of
individuals over a number of years and are useful to
provide informaon about disease progression and other
key factors for disease and survival. They are generally
completely researcher driven but, nonetheless, can be
useful for assessing populaon health. Studies in the
United States that have tracked individuals include the
Framingham cohort, the Nurses Health Study and the
American Cancer Societys Cancer Prevenon Study.
Health facility assessments
Health facility assessments are intended to capture the
resources and inputs of a specic health centre, be it
a primary care clinic, a community health centre or a
specialty clinic. Modules include facility infrastructure,
health centre budget reviews, pharmaceucal
inventories, secondary output review, and services
for specic condions such as tuberculosis treatment.
As a result, it is advisable to regularly assess the quality
of health service data and to help ensure some basic
standardisaon, to the extent possible, to beer serve
naonal and regional interests. Regular monitoring also
helps to beer understand the aggregate capacity of a
health system to provide care. Supervisory systems can
be used to collect standardised and systemac data and
to provide comparisons over me and between clinics
and regions. Addional data may be collected through a
health facility survey, which is usually based on a sample
of clinics.
Such a survey may consider dierent aspects of service
quality such as the availability of drugs, commodiesand trained sta. Special techniques such as record
review or observing clientprovider interacons can
add considerable value to the assessment, but they also
increase costs and complexity. Data collected from record
reviews and stang inventories can be used to validate
roune administrave stascs on the volume of services
delivered and on the availability and geographical
distribuon of human resources.
Census data
Where available, populaon-level census data can serve
as the primary informaon source for determining the
size of a populaon; its geographical distribuon; and
the social, demographic and economic characteriscs
of its people. Censuses have been undertaken in most
countries in recent decades and, in some places, for
more than a century. The Stascs Division of the United
Naons Department of Economic and Stascal Aairs
(UNDESA) has developed principles, recommendaons
and manuals for populaon and housing censuses
available from their website (UNDESA 2010).
From a health perspecve, informaon on populaonnumbers and distribuon by age, sex and other
characteriscs is essenal for naonal and local planning,
esmang target populaon sizes and trends, and
evaluang rates of service coverage and future needs.
Census data can also provide valuable informaon
on some key health outcomes, parcularly mortality.
Informaon on major health determinants and other
key factors such as poverty, housing condions, water
and sanitaon, can also be collected in a census. The
nature of the census allows for small-area esmaon and
disaggregaon by key straers such as socioeconomic
status. Censuses can also provide valuable informaonon the number of health professionals working in the
health sector.
-
8/10/2019 Improving the Quality and Use
15/20
13Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
Health facility assessments are important in evaluang
not only the resources that are necessary to provide
health services to a populaon, but also to evaluate the
quality of the services being provided and factors related
to the regular provision of medicaments, such as stock-
out rates.
DHS Macro conducts service provision assessments in
selected countries on a quasi-regular basis, gathering
informaon from health facilies on the type and quality
of care that they provide. The informaon provided to
policy-makers includes data on health facilies in-country,
and their resources, and basic systems and specic
health services (eg basic child health services or maternalcare).
Secondary macro-datasets
Aggregate health indicators, such as those created by
internaonal organisaons, are generally based on
datasets that aggregate individual data. They are usually
a mixture of datasets, oen poorly documented. Two
major problems exist with these types of datasets.
First, some countries do not have data on the quanty
of interest. Therefore, although some aggregates can
be compared, they are not enrely comparable cross-
naonally, due to missing informaon from some
countries. Second, those who create these aggregates
(eg WHO, World Bank, United Naons Childrens Fund,
United Naons Development Programme and United
Naons Populaon Fund) do not always detail how the
aggregate was generated. What is worse, countries oen
report back to WHO the esmates that WHO provides as
being their naonal esmates. This pracce discourages
countries investment and interest in developing naonal
health informaon systems designed to meet theirhealth development needs.
Data that are generated through research are oen
coupled with results that are very important for policy.
These sources vary widely, depending on the aim of the
research, and therefore the data obtained are dicult to
categorise. However, they are useful as further sources
of informaon. Thus far, the availability of these datasets
has not been standardised, but there is increasing
pressure to ensure that such data be made available
by publicaon. The Instute for Health Metrics and
Evaluaon (IHME), for example, has included this as a
core principle of its operaon, and makes all data used in
publicaon available on its website (IHME 2010).
-
8/10/2019 Improving the Quality and Use
16/20
-
8/10/2019 Improving the Quality and Use
17/20
15Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowled
geHub
Lozano R, Solis P, Gakidou E, et al (2006). Benchmarking
performance of Mexican states with eecve coverage.
Lancet 368(9548):17291741.
Mokdad AH, Stroup DF and Giles WH (2003). Public
health surveillance for behavioral risk factors in a
changing environment. Recommendaons from the
Behavioral Risk Factor Surveillance Team. Morbidity and
Mortality Weekly Report. Recommendaons and Reports
52(RR-9):112.
Murray CJL and Frenk J (2000). A framework for
assessing the performance of health systems. Bullen of
the World Health Organizaon79(6):717732.
Murray CLJ, Lopez A, Barofsky J et al (2007). Esmang
populaon cause-specic mortality fracons from in-
hospital mortality: Validaon of a new method. PloS
Medicine4(11):e326 0112.
OECD (Organisaon for Economic Co-operaon and
Development) (2000).A System of Health Accounts,
OECD, Paris.
PAHO (Pan American Health Organizaon) (2001).
Indicadores de Salud: Elementos bsicos para el anlisis
de la situacin de salud.Bolen Epidemiolgico4:15.
PARIS21 (Partnership in Stascs for Development in the
21st Century) Secretariat.A Guide to Designing a Naonal
Strategy for the Development of Stascs(NSDS).www.
paris21.org/sites/default/les/1401_0.pdf (Accessed
18 November 2011).
PHDSC (Public Health Data Standards Consorum)
(2007). Building a Roadmap for Health Informaon
Systems Interoperability for Public Health (Public
Health Uses of Electronic Health Record Data).www.
ihe.net/Technical_Framework/upload/IHE_PHDSC_Public_Health_White_Paper_2007_10_11.pdf (Accessed
1 March 2010)
Poullier J, Hernndez P and Kawabata K (2003).
Naonal Health Accounts: Concepts, Data Sources, and
Methodology. Health Systems Performance Assessment:
Debates, Methods and Empiricism, World Health
Organizaon, Geneva.
RIPSA (Interagency Network for Health Informaon)
(2002).Indicadores Bsicos Para a Sade No Brasil:
Cenceitos e Aplicacoes, RIPSA, Brasilia.
AbouZahr C, Adjei S and Kanchanachitra C (2007). From
data to policy: Good pracces and cauonary tales.
Lancet369:10391046.
Aqil A, Lippeveld T and Dozumi D (2009). PRISM
framework: a paradigm shi for designing, strengthening
and evaluang roune health informaon systems.
Health Policy and Planning 112.
Ash J, Berg M and Coiera E (2004). Some unintended
consequences of informaon technology in health care:
the nature of paent care informaon system related
errors.Journal of the American Medical Informaon
Associaon11:104112.
Detmer D (2003). Building the naonal health
informaon infrastructure for personal health, health
care services, public health and research. BMC Medical
Informacs3:112.
Evans T and Stanseld S (2003). Health informaon in
the new millennium: a gathering storm? Bullen of the
World Health Organizaon81(12):856.
Haux R (2006). Health informaon systemspast,
present, future.Internaonal Journal of Medical
Informacs75:268281.
HIS Hub (Health Informaon Systems Knowledge Hub).
www.uq.edu.au/hishub/about-the-his-hub (Accessed
March 1 2010).
HMN (Health Metrics Network) (2008).Assessing the
Naonal Health Informaon System: An Asssessment
Tool (Version 4.00), World Health Organizaon, Geneva.
IHME (Instute for Health Metrics and Evaluaon).
www.healthmetricsandevaluaon.org (Accessed
1 March 2010).
Krickeberg K (2007). Principles of health informaon
systems in developing countries.Health Informaon
Management Journal36(3):820.
Lenz R, Elnstner T, Siegele H et al (2002). A praccal
approach to process support in health informaon
systems.Journal of the American Medical Informaon
Associaon9:571585.
Lozano, R (2008). Es posible seguir mejorando los
registros de las defunciones en Mxico?Gaceta Medica
de Mxico144(6):525533.
References
-
8/10/2019 Improving the Quality and Use
18/20
16 Working Paper Series Number 5 November 2009
HealthInformaonSystemsKnowledg
eHub
Secretara de Salud (2002).Informacin para la rendicin
de cuentas 2001, Federal Government, Salud, Mxico.
WHO (World Health Organizaon) (1993). Guidelines for
the Development of Health Management Informaon
Systems, WHO, Geneva. www.wpro.who.int/publicaons/
pub_9290611065.htm (Accessed 22 November 2011).
WHO (World Health Organizaon) (2000). The World
Health Report 2000Health Systems: Improving
Performance, WHO, Geneva.
WHO (World Health Organizaon) (2007). Everybodys
Business: Strengthening Health Systems to ImproveHealth Outcomes, WHOs Framework for Acon,
WHO, Geneva.
World Bank, WHO (World Health Organizaon)
and USAID (United States Agency for Internaonal
Development) (2002). Guide to Producing NHA with
Special Applicaons for Low-Income and Middle-Income
Countries. World Bank, WHO and USAID, Washington, DC.
UNDESA (United Naons Department of Economic and
Stascal Aairs). 2010 World Populaon and Housing
Census Programme. hp://unstats.un.org/unsd/
demographic/sources/cwp2010/docs.htm (Accessed1 March 2010).
-
8/10/2019 Improving the Quality and Use
19/20
The Knowledge Hubs for Health Iniave
The Health Informaon Systems Knowledge
Hub is one of four hubs established by
AusAID in 2008 as part of the Australian
Governments commitment to meeng
the Millennium Development Goals and
improving health in the Asia and Pacic
regions. All four hubs share the common
goal of expanding the experse and
knowledge base to help inform and guide
health policy.
The Knowledge Hubs are funded by
AusAIDs Strategic Partnership for
Health Iniave.
Health Informaon Systems Knowledge Hub
The University of Queensland
Aims to facilitate the development and integraon of health
informaon systems into the broader health system strengthening
agenda, and increase local capacity to ensure that cost-eecve,
mely, reliable and relevant informaon is available. The Health
Informaon Systems Knowledge Hub also aims to beer inform
health informaon systems policies across Asia and the Pacic.
www.uq.edu.au/hishub
Human Resources for Health Knowledge Hub
The University of New South Wales
Aims to contribute to the quality and eecveness of Australias
engagement in the health sector in the AsiaPacic region by
developing innovave policy opons for strengthening human
resources for health systems. The hub supports regional, naonal
and internaonal partners to develop eecve evidence-informed
naonal policy-making in the eld of human resources for health.
www.hrhhub.unsw.edu.au
Health Finance and Health Policy Knowledge Hub
The Nossal Instute for Global Health(University of Melbourne)
Aims to support regional, naonal and internaonal partners
to develop eecve evidence-informed naonal policy-making,
parcularly in the eld of health nance and health systems. Key
themac areas for this hub include comparave analysis of health
nance intervenons and health system outcomes; the role of
non-state providers of health care; and health policy development
in the Pacic.
www.ni.unimelb.edu.au
Compass: Womens and Childrens Health Knowledge Hub
Compass is a partnership between the Centre for Internaonal
Child Health, The University of Melbourne, Menzies School
of Health Research and Burnet Instutes Centre for
Internaonal Health.
Aims to enhance the quality and eecveness of women's and
childrens health intervenons and focuses on supporng the
Millennium Development Goals 4 and 5improved maternal
and child health, and universal access to reproducve health. Key
themac areas for this hub include regional strategies for child
survival; strengthening health systems for maternal and newborn
health; adolescent reproducve health; and nutrion.
www.wchknowledgehub.com.au
-
8/10/2019 Improving the Quality and Use
20/20
A strategic partnerships iniave funded by the Australian Agency for Internaonal Development