Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
12
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Towards a comprehensive measurement
of the health system in Latin America:
presentation of the multidimensional
performance index
Nicolás Rubio García
1
Jairo Alfredo Ortiz Pacheco
2
Andrés Mateo Gómez Franco
3
To reference this article / Cómo citar este artículo / Para citar
este artigo:
Rubio-García, N., Ortiz-Pacheco, J. A., & Gómez-Franco,
A. M. (2024). Towards a comprehensive measurement of the healthy
system in Latin America: presentation of the multidimensional
performance index.
Revista Criterios, 31
(2), 12-27.
https://doi.
org/10.31948/rc.v31i2.4000
Received:
March 29, 2024
Revised:
July 4, 2024
Accepted:
August 3, 2024
Abstract
An innovative methodology for measuring health system performance in Latin
America using a multidimensional approach is presented. The methodology is
based on the weighting of four key sub-indices: coverage, quality, socioeconomic
health, and population perception, which together form the Multidimensional
Performance Index (MDI). The main objective is to provide a comprehensive
view of health system performance in the region and to identify potential areas
for improvement. The research is aligned with the World Health Organization’s
(WHO) defnitions of health as a holistic state of well-being and recognizes its
multidimensional nature. In this framework, the health system is understood
as an integral structure of organizations, institutions, and resources dedicated
to improving health in general. The proposed methodology includes three main
approaches: comparative, predictive, and evaluative, essentially aimed at
improving the performance of the health system by facilitating international and
national comparisons, predicting future behavior and identifying the variables
that infuence the main indexes. Finally, the methodology is applied to evaluate
the performance of 18 Latin American countries, with a specifc comparative
analysis in three of them: Colombia, Costa Rica, and Dominican Republic.
Keywords:
health; welfare; health statistics; social security; health service;
health policy
1
Student of Specialization in Visual Analytics, Fundación Universitaria Internacional de la Rioja and Big Data; political scientist,
Universidad Nacional de Colombia, Medellín; student of Economics and Finance, Universidad El Bosque. E-mail address: nrubiog@
unal.edu.co
2
Law student, Universidad de Antioquia; Political Science student, Universidad Nacional de Colombia, Medellín. E-mail address:
jortizpa@unal.edu.co
3
Student of Political Science, Universidad Nacional de Colombia, Medellín. E-mail address: angomezf@unal.edu.co
This article is the result of the paper presented at the VII National Congress of Political Science Uniandes - ACCPOL, on June 9, 2023,
and of the fnal work of the subject Comparative Politics.
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
13
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Hacia una medición integral del
sistema de salud en América
Latina: presentación del índice de
desempeño multidimensional
Resumen
Se presenta una metodología innovadora para medir el desempeño del sistema de
salud en América Latina mediante un enfoque multidimensional. La metodología
se basa en la ponderación de cuatro subíndices clave: cobertura, calidad, salud
socioeconómica y percepción de la población, que en conjunto conforman el índice
de desempeño multidimensional (IDM). El objetivo principal es proporcionar una
visión integral del desempeño del sistema de salud en la región e identifcar
áreas potenciales de mejora. La investigación está alineada con las defniciones
de la Organización Mundial de la Salud (OMS), que concibe la salud como un
estado de bienestar holístico y reconoce su naturaleza multidimensional. En
este marco, el sistema de salud se entiende como una estructura integral que
abarca organizaciones, instituciones y recursos dedicados a la mejora de la
salud en general. La metodología propuesta abarca tres enfoques principales:
comparativo, predictivo y evaluativo. Estos enfoques tienen como objetivo mejorar
el desempeño del sistema de salud al facilitar la comparación internacional y
nacional, prever el comportamiento futuro e identifcar las variables que infuyen
en los índices principales. Finalmente, la metodología se aplica para evaluar
el desempeño de 18 países latinoamericanos, con un análisis comparativo
específco en tres países: Colombia, Costa Rica y República Dominicana.
Palabras clave
: salud; bienestar; estadísticas sanitarias; seguridad social;
servicio de salud; política de la salud
Rumo a uma medição abrangente
do sistema de saúde na América
Latina: apresentação do índice de
desempenho multidimensional
Resumo
É apresentada uma metodologia inovadora para medir o desempenho do sistema
de saúde na América Latina usando uma abordagem multidimensional. A
metodologia é baseada na ponderação de quatro subíndices principais: cobertura,
qualidade, saúde socioeconômica e percepção da população, que formam o
Índice de Desempenho Multidimensional. O principal objetivo é fornecer uma
visão abrangente do desempenho do sistema de saúde na região e identifcar
possíveis áreas de melhoria. A pesquisa está alinhada com as defnições de
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
14
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
saúde da Organização Mundial da Saúde (OMS) como um estado holístico de
bem-estar e reconhece sua natureza multidimensional. Nessa estrutura, o
sistema de saúde é entendido como uma estrutura integral de organizações,
instituições e recursos dedicados à melhoria da saúde em geral. A metodologia
proposta inclui três abordagens principais: comparativa, preditiva e avaliativa.
Essas abordagens têm como objetivo melhorar o desempenho do sistema
de saúde, facilitando as comparações internacionais e nacionais, prevendo o
comportamento futuro e identifcando as variáveis que infuenciam os principais
índices. Finalmente, a metodologia é aplicada para avaliar o desempenho de
18 países latino-americanos, com uma análise comparativa específca em três
deles: Colômbia, Costa Rica e República Dominicana.
Palavras-chave:
saúde; bem-estar; estatísticas de saúde; seguridade social;
serviço de saúde; política de saúde
Introduction
The right to health encompasses not only the
absence of disease, but also the conditions
necessary for a dignifed life. To guarantee this
right, states establish institutions to administer
and manage the health system. The World Health
Organization (WHO) defnes this system as the
set of organizations, institutions, and resources
that aim to improve the health of the population
and that require for their operation adequate
fnancial, human, and material resources to
provide treatments and services that meet the
needs of the population (Ocampo-Rodríguez
et al., 2013). Therefore, although the health
system is managed by government entities, it
is composed of a series of organizations with
the common objective of improving the health
of the population.
The collective efort of these organizations
shapes the health system and determines
its characteristics according to the needs of
the population, which vary from country to
country and have general characteristics such
as coverage, quality of services, socioeconomic
factors and public perception. In this sense, it
is essential to identify its shortcomings in order
to improve its efciency (Bernal and Gutiérrez,
2012; Rodríguez, 2019). Likewise, recognizing
the qualities of an efcient health system makes
it possible to replicate them to provide standards
that meet the needs of the population.
In this context, on July 22, 1946, the Constitution
of the World Health Organization (WHO,
2022) was signed by 61 States, introducing a
multidimensional approach to health, defned
as a complete state of physical, mental and
social well-being, and establishing global health
as essential for peace and security. Achieving
this requires the cooperation of individuals and
states through the adoption of appropriate
health and social policies. Under this premise,
the concept of health has evolved over time,
refecting diferent approaches. To provide a
comprehensive understanding, two perspectives
are explored: historical and evolutionary.
From a historical perspective, health has had
three diferent conceptions: traditional, modern
and postmodern (Marín et al., 2020; Serrano et
al., 2019; Vergara, 2007). The frst moved from
the health-disease dichotomy to Hippocratic
naturalism, moving away from supernatural
perceptions and focusing on the biological study
of diseases and the human body. Over time,
economic, political, and social aspects and their
relationship to health gained importance, giving
rise to the modern conception that incorporated
health studies into social representations.
Postmodern conceptions, in turn, emphasized
the dynamic nature of health, integrating
biological advances and other knowledge
systems and recognizing the need for health to
adapt to social agreements about its content.
From an evolutionary perspective, the concept
of health initially focused on the health-disease
polarity, defned as the absence of disease.
Health then came to be understood as a state
of normality of organic and psychic functions,
i.e. a state of harmony and balance with the
organism. This evolution culminated in a
concept of health as complete well-being in its
physical, psychological, and spiritual aspects.
Based on these concepts, health has adopted
an integrative approach that emphasizes the
individual’s capacity to develop his or her
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
15
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
personality and functioning. Thus, health is
understood as an integral and holistic state
in which individuals can experience free and
complete development. It thus recognizes the
interrelationship with the social environment
and emphasizes the importance of an adequate
environment, access to appropriate health
services, and favorable social and cultural
conditions to ensure the integral development
of the individual.
The ability to guarantee this comprehensive
nature of health depends to a large extent on
the social security system, which includes the
measures and mechanisms implemented by
a State to protect its citizens against various
risks. Therefore, social security is a component
of a public right that is closely linked to the
constitutional principles of each State (Álvarez
et al., 2021), whose main objective is to enable
people within its jurisdiction to claim their rights
in the best way possible, while ensuring the
efciency and efectiveness of its procedures
(Bernal and Gutiérrez, 2012).
This study focused on the health aspect of
the general social security system, in which
the method of fnancing the system is an
essential factor in determining its functioning
and characteristics. According to Bernal and
Gutiérrez (2012), there have traditionally been
three health fnancing systems: the private,
which includes formal or informal insurance
arrangements; the public, which depends on
resources from the national budget; and fnally,
the contributory social security systems, which
generally receive contributions from individuals,
often linked to their employment status.
On the other hand, several studies have
analyzed how socioeconomic inequalities can
afect the management of health systems. For
example, poverty and social marginalization
may limit people’s access to health services
and their ability to meet minimum quality of
life needs. This concept is known as the social
determinants of health (Avilés, 2017; Buitrago
et al., 2023; Fernández et al., 2019; Gil de
Miguel and Campuzano, 2014; London et al.,
2009; Sinchi, 2020). According to this idea,
health promotion is responsible for studying
and intervening in these social, environmental
and economic conditions to mitigate their
impact on public and individual health, as well
as the positive determinants of health that
contribute to improving people’s quality of life
(De la Guardia and Ruvalcaba, 2020).
In summary, the analysis and study of the
health system has often been approached from
economic, social security, inequality, and clinical
perspectives. However, these perspectives are
often studied in isolation, neglecting the links
and infuences between them. Consequently,
there is a need for an integrative analysis that
incorporates multiple variables and perspectives
to provide a comprehensive understanding of
health system behavior. This holistic approach
would provide insight into the complex dynamics
of the system, which in turn would lead to more
efective policies and interventions. Therefore,
it is decided to address this issue from the
perspective of system performance.
Some contemporary approaches suggest that
performance measurement serves multiple
purposes, such as evaluation, monitoring,
budgeting, motivation, promotion, celebration,
learning, and improvement (Agasisti et al.,
2020; Behn, 2003). In this sense, stakeholders
can improve accountability and transparency
by adopting a holistic approach to performance
measurement. Moreover, in the public sector, it
allows citizens to evaluate the efectiveness of
government and make informed judgments, as
well as empowering ofcials by providing them
with the necessary information to improve
performance and allocate resources efectively
(Carrillo et al., 2021). Therefore, measuring
performance allows understanding its behavior,
promotes accountability and transparency,
facilitates decision-making, and promotes
continuous improvement (Behn, 2003; Vera et
al., 2022).
Accordingly, a methodology has been
developed to measure the performance of
the health system in Latin America from a
multidimensional perspective, based on an
approach that considers the main perspectives
of the study of health, through four sub-indices
that are integrated into a main tool that has
been called the Multidimensional Performance
Index (MDI).
Thus, this study not only contributes to a better
understanding of the functioning and impact
of health systems in Latin America, but also
highlights the importance of measuring their
performance. By examining the strengths,
weaknesses, and characteristics of health
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
16
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
systems and their operating contexts, the
objective is to shed light on how they infuence
service delivery and impact on the well-being of
the population, specifcally in terms of coverage,
quality, socioeconomic health, and perception.
Based on the information provided, the
hypothesis proposed is that the sub-indices that
make up the multidimensional performance
index can be used to explore the future
behavior of health systems, compare systems
at the national or regional level, and identify
the factors that infuence each system.
Methodology
A mixed methodology was used, integrating
follow-up techniques such as document review
and bibliographic searches in databases such as
Dialnet, Scielo and Elsevier, as well as tracking
bibliographic similarities between references
using Litmaps and searching international
and regional statistical repositories of
organizations such as ECLAC, WHO, PAHO and
Latinobarómetro.
The objective was to select indicators with at
least three or more observations in 18 Latin
American countries
4
during the period 2011-
2019. In cases where data were missing,
Multivariate Imputation by Chained Equations
(MICE) techniques and sensitivity measures
5
were used to estimate missing values. It was
also decided to standardize the data on a scale
from 0 to 1, using the min-max method, in
order to ensure a uniform scale and guarantee
the equal contribution of each indicator.
The analysis included the collection of
30 indicators in diferent areas, ranging
from demographic variables such as total
population, to socio-economic variables such
as the proportion of people living in poverty,
the net enrollment rate in high school, and the
employed population. It also included economic
aspects such as gross domestic product (GDP),
health expenditures, epidemiological indicators
(malnutrition rate, incidence of communicable
and non-communicable diseases, morbidity
rates by specifc groups), and coverage of
maternal, neonatal, and reproductive health
services.
4
Cuba and Haiti were not included due to lack of information.
Neither were Guyana, Suriname and French Guiana.
5
Relative bias, root mean square error (RMSE),
.
Based on these aspects, four additional
indicators were created to complement the
analysis. These 34 indicators were used
to construct four sub-indices representing
diferent dimensions of the health system.
Then, by weighting the four sub-indices using
the geometric mean, due to its efectiveness in
handling percentages and its lower sensitivity to
extreme values than the arithmetic mean, the
Multidimensional Performance Index (MDI) was
developed, which provides a comprehensive
measure of the health system.
In the results section, an evaluative approach
to the methodology was applied, analyzing the
relationship between the exogenous variables
and the MDI through regression models with a
5% (
) signifcance level.
Interpretation of the Multidimensional
Performance Index
The methodology is designed in such a way
that values close to 1 indicate a high level in
the respective dimension, while values close to
0 indicate the opposite.
The frst index (
) represents the health
coverage of the country’s population. Its
measurement is based on SDG 3.8.1 on
Universal Health Coverage (WHO, 2022),
except that
Capacité
is replaced by PCS in
the weighting. This provides an overview of
health coverage in a country, considering the
following variables:
−
Insured population (PCS) represents the
proportion of the population with some
form of health insurance in the country.
−
INF: The coverage of infectious diseases
in the country.
−
NCDs: Coverage of non-communicable
diseases, which are chronic conditions
that often require long-term treatment
and care.
−
RMNCH: Coverage of reproductive,
maternal, newborn and child health.
Focuses on critical areas of women’s and
children’s health.
(1)
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
17
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
The second index (
) focuses on modeling
the quality of a health system by assessing
its capacity to respond to and efectively
control adverse health factors, considering
the availability of human
6
(
) and physical
7
(
) resources. Both are multiplied by the
percentage of gross domestic product invested
in health (
). To invert the scale, 1 is
subtracted from each negative factor, i.e.:
−
Adjusted mortality rate (AMR): the
diferences in mortality rates between
populations of diferent ages and sexes
relative to the crude mortality rate (CMR).
−
Undernutrition (SoU): the proportion of the
population that does not have sufcient
access to nutritious food or regularly
consumes inadequate amounts of food.
−
Morbidity (IT): The incidence of diseases
and health problems afecting the
population, such as tuberculosis, malaria,
HIV, tropical diseases, and stunted and
overweight children.
(2)
The third index (
focuses on the health and
well-being of the population by incorporating
factors related to the social determinants of
health. The components are as follows:
−
Net secondary school enrollment (Ab): The
percentage of the school-age population
enrolled at this level.
−
Population with health insurance (PCS).
−
Labor force participation rate (LFP): The
percentage of the working-age population
(WAP) that is in the labor force.
−
Impact of poverty on health (POB or
TBM): The negative impact of poverty on
health, considering the population living in
poverty and the crude mortality rate.
−
Adjusted life expectancy (ALE): the
number of years a person can expect to
live (LE), bearing in mind the impact of
poverty on health.
6
Summing the available data for four types of professionals:
medical, dental, nursing, and pharmacy staf.
7
Number of hospital beds per thousand inhabitants.
The choice of the net enrolment rate in
secondary education instead of the often-used
literacy rate is due to the fact that it guarantees
a basic education in science, mathematics and
social sciences, while the literacy rate considers
the population that can read and write after a
certain age, but not the felds of knowledge
in which they have been educated. On the
other hand, it is signifcant to note that the
employment rate does not distinguish between
formal and informal employment conditions.
Despite these limitations, the rate provides
valuable information on the labor market and
its possible impact on the health system.
(3)
The fourth index (
) measures the population’s
perception of health and related issues through
the question: What do you consider to be
the most important problem in the country?
from the Latinobarómetro regional survey,
representing 100% of the total population
8
.
The relative frequency of responses is analyzed
for the following answers: Health Problems (
), Poverty (
), Crime - Public Safety (
) and Unemployment - Joblessness (
).
It is important to note that most countries
do not have their own public opinion surveys
on health, and when they do exist, they are
usually not conducted annually; therefore, the
Latinobarómetro is the closest available tool
for consultation. It is also noteworthy that the
percentage of respondents choosing
Health
problems
as the most signifcant issue does not
exceed 5%
9
. It is so necessary to include other
perceptions related to general well-being.
In the index, the highest weight is given
to
, as this is the most direct indicator
of the population’s perception of health. It
also refects the insured population, whose
perceptions may be infuenced by availability
and accessibility. The inclusion of the three
additional perception categories provides a
more complete understanding of how the
population perceives health. This integration
is crucial because health cannot be separated
from other aspects of an individual’s life (De
La Guardia and Ruvalcaba, 2020; Peres et al.,
8
Except Panama, which is between 98% and 100%.
9
Except in Brazil, where it is the main problem in most years
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
18
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
2023). It is worth noting that even if people
do not predominantly choose
Health problems
,
this does not mean that they do not consider
them to be a problem. Rather, they may
perceive other issues to be of greater concern.
(4)
The main index (
) is the health system
performance or MDI. It combines the four sub-
indices discussed above and weights them
equally using the geometric mean to reduce
sensitivity to extreme values and maintain
relationships of magnitude. A value close
to 1 would indicate that the health system
performance is optimal, with high coverage, high
quality of health services, healthy population
and positive public opinion about the system.
(5)
Results
The WDI was used to analyze the performance
of 18 countries for the period 2011-2019, which
allowed the identifcation of the country with
the highest and lowest performance, as well
as the average WDI, as an approximation of
the performance of the Latin American region
and, consequently, as a threshold to defne the
best performing countries and those that could
improve. In this context, Table 1 shows the
simple average WDI of the 18 countries over a
nine-year period:
Table 1
Simple average of the Multidimensional
Performance Index of the 18 countries (2011-
2019)
Country
Costa Rica
73,43 %
Chile
70,89 %
Uruguay
70,42 %
Colombia
69,05 %
Panamá
63,52 %
Ecuador
63,40 %
Nicaragua
62,82 %
Perú
61,89 %
Argentina
61,62 %
México
61,48 %
Brasil
61,14 %
Latin America
59,04 %
Paraguay
58,05 %
Guatemala
56,44 %
El Salvador
56,31 %
Honduras
52,14 %
Venezuela
51,81 %
Bolivia
46,22 %
Dominican Republic
44,02 %
Note.
The yellow color refers to the average for
the region; and, the green and red colors are,
respectively, the countries above and below
the average.
The results showed that the regional average
was 59%, which implies a large margin for
improvement in the diferent dimensions of
the population health index. The country with
the highest MDI was Costa Rica with 73.4%,
followed by Chile with 70.9% and Uruguay with
70.4%. Conversely, the Dominican Republic was
the country with the lowest score on the index,
with 44%, a large percentage diference from
Costa Rica. Finally, Colombia was in the group
of countries with a medium performance; it
was even close to Chile and Uruguay by about
two percentage points.
Against this background, it was interesting to
compare the WDIs of Costa Rica, Colombia
and the Dominican Republic in order to identify
the contributions of diferent variables or
factors to their performance and to examine
individual trends over the period. However, as
in this work it was essential to recognize the
need for a multidimensional analysis when
assessing the performance of these health
systems, it was necessary to show that a
single factor analysis was insufcient, even
when seeing an aggravating factor common
to the region, such as poverty, which acts
as a barrier preventing access to essential
services and ultimately afects overall well-
being (Fernandez et al., 2019).
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
19
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
In Colombia, it was vital to highlight the alarming rates of poverty and extreme poverty that
prevailed during the frst decade of the century. In 2002, 74.6% of the national population lived in
such conditions. However, there was a gradual decline throughout the decade, reaching a rate of
57% in 2010. Despite this 17.6% reduction, it was crucial to recognize that more than half of the
population remained in conditions of vulnerability. This reality signifcantly impeded their access to
basic services such as basic sanitation, reliable energy and drinking water services, sufcient food,
adequate security and quality education. As a result, these challenges made it almost impossible
to achieve full well-being, including optimal health outcomes (Avilés, 2017; Buitrago et al., 2023;
Fernández et al., 2019).
In Figure 1, the WDI for Colombia showed a performance ranging from 67.4% to 70.2%, with
the lowest point in 2012 and the highest point in 2016. It is worth noting that over the years,
performance has shown an upward trend relative to the baseline. This suggests that, although the
peak performance was not reached in 2019, the level of performance did not decline sufciently to
return to the baseline. This indicates that there has been a sustained improvement in basic service
coverage, system quality, socio-economic health and population perceptions in Colombia.
Figure 1
Colombia WDI time series
It is worth noting that Colombia’s better performance, despite having a higher proportion of people
living in poverty than Latin America, can be explained within the WDI framework, as it integrates
multiple dimensions and provides a comprehensive perspective that recognizes the infuence
of diferent approaches. In order to identify the factors infuencing performance, a regression
analysis was carried out to examine the relationship between a set of explanatory variables and
the dependent variable.
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
20
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Table 2
Regression results using MDI as criterion for Colombia
Predictor
bb
IC of 95%
SR^2(>|t|)
Fit
(Intercept)0.49*[0.30, 0.68]0.00385*
R
2
= 0,967*
Poly (PCS, 2)10.02[-0.00, 0.04].110.05153IC 95% [.09,.97]
Poly (PCS, 2)2-0.01[-0.02, 0.00].070.08481
Ab0.25[-0.02, 0.53].100.06049
d
s
2
-0.00[-0.00, 0.00].100.05862
GTS0.07[-2.03, 2.17].000.9263
Note.
b represents unstandardized regression
weights. SR^2, the squared semi partial
correlation. LL and UL indicate the lower
and upper limits of a confdence interval,
respectively. * indicates p < .05.
The regression model used included four
predictor variables together with the intercept,
which represents the possible value of the
dependent variable if there is no infuence
from the predictor variables, which was found
to be statistically signifcant. However, three
variables had p-values close to 0.05.
An interesting aspect to consider is the joint
efect of these variables on the WDI. The
coefcients of GTS, Ab and PCS were positive,
suggesting that an increase in their magnitudes
contributes to an increase in WDI. However, a
signifcant increase in
and PCS can lead to
a decrease in WDI. This can be explained by
the assumption that an excessive increase in
the number of insured persons may afect the
ability of the system to provide health services.
Similarly, a signifcant increase in the health
workforce without the necessary funding and
infrastructure to support a larger workforce
may have negative consequences.
On the other hand, Ab was found to be benefcial,
indicating that there is room for improvement in
this area. Finally, the regression model showed
a high and signifcant level of explanatory
power: 96.7%, and when adjusted, it explains
91.2% of the variation.
In contrast, Costa Rica did not have such high
levels of poverty and extreme poverty in the
2000s. In 2002, for example, the percentage
of the population living in these conditions was
33.4%, a fgure that fell to 23.1% in 2010. It
should be noted that Costa Rica’s population
living in poverty was lower than the regional
average: in 2002, the regional average was
57.5%, while in 2010 it was 40.2% (Economic
Commission for Latin America and the
Caribbean [ECLAC], 2022).
Figure 2 shows the WDI for Costa Rica, which
fuctuated between 74.7% and 72.6%, with the
highest point observed in 2011 and the lowest
in 2014. It is worth noting that during the
period analyzed, the WDI experienced a slight
decline of around 1 percentage point compared
to the baseline.
To understand the reasons for this decline,
it was necessary to examine the four sub-
indices. The analysis showed that the quality
index decreased by 8% and the perception
index by 2%. Despite these declines, both sub-
indices were relatively high compared to other
Latin American countries. Consequently, the
question arises as to which factors can explain
the variation in the WDI for Costa Rica, using
regression analysis.
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
21
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Figure 2
Time series of the WDI for Costa Rica
Table 3
Regression results using MDI as a criterion for Costa Rica
Predictor
b
b
IC of
95%
SR^2(>|t|)
Fit
(Intercept)0.21[-0.03, 0.46]0.06888
R
2
= 0,953*
EV0.42*[0.17, 0.67].270.00896*IC del 95 % [.34,.96]
Poly (GTS, 2)10.01*[0.01, 0.02].490.00296*
Poly (GTS, 2)20.00[-0.00, 0.01].020.22656
NCD0.32*[0.15, 0.49].330.00601*
Note.
b represents unstandardized regression weights. SR^2 signifes the squared semi-partial
correlation. LL and UL indicate the lower and upper limits of a confdence interval, respectively. *
specifes p < .05.
The regression model for Costa Rica included three signifcant predictor variables with positive
coefcients. To better understand these results, the following explanation was proposed: frst,
the decline in the fertility rate associated with the ageing of the population had an impact on the
health status of the population. It was therefore reasonable to assume that an increase in life
expectancy would have a positive impact on the performance of the health system.
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
22
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Second, performance was expected to improve as NCD coverage increased, but it was interesting
to note that an increase in STG had only a modest impact on performance. Moreover, even
when spending increased substantially, the improvement in performance was comparatively less
than when spending was lower. This suggests that the management and administration of public
resources for health in the country was sufciently optimized that a signifcant increase in spending
would not necessarily result in an equivalent increase in welfare.
In terms of variation, the regression model explained 95.3% of the total variation and, when
adjusted, 90.5% of the variation.
The Dominican Republic experienced a signifcant increase in the proportion of its population
living in poverty and extreme poverty over the period analyzed. In 2002, 45.1% of the population
lived in these conditions, a fgure lower than that for Latin America, which was 57.5%. By 2010,
however, the proportion of the population living in these conditions had increased signifcantly to
53.2 per cent, compared with the regional average of 40.2 per cent.
Figure 3
Dominican Republic, WDI time series
In the case of the Dominican Republic, the WDI
values fuctuated between 42.5% and 44.7%,
with a high in 2011 and a low in 2015. However,
these fuctuations were relatively small, difering
by only a few decimal points. What is worrying
is that the peak is around 15 percentage points
below the average, indicating a signifcant
divergence with the region. Interestingly, the
decline in the population living in poverty
between 2011 (50.8%) and 2019 (22.9%) did
not correspond to a signifcant improvement
in the WDI. This raises questions about the
factors infuencing the overall performance.
When analyzing the sub-indices, three of them
were in the range of 50% to 74%, indicating a
moderate level of performance. Nevertheless,
the quality index showed worryingly low values
of between 14% and 16%. This was mainly
due to the constant lack of observations for
two variables:
and
, which required
the use of imputation methods. It should be
noted that these methods have also been
used in other countries without having a
signifcant impact on the sub-indices. In
addition, sensitivity tests were carried out
for all countries and the results were within
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
23
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
acceptable limits for all three measures. Therefore, in order to fully understand the cross-
country variation in the WDI, further analysis was needed to investigate and explore the factors
contributing to these results.
Table 4
Regression results using MDI as the criterion for Dominican Republic
Predictorbb IC of 95%SR^2(>|t|)Fit
(Intercept)1.23*[0.32, 2.13]0.0196*
R
2
= 0,915*
TMA-0.54*[-1.03, -0.06].200.03625*IC 95% [.05,.93]
Poly (PNS, 2)1-0.03*[-0.04, -0.01].550.00713*
Poly (PNS, 2)2-0.00[-0.01, 0.01].030.31639
Log (GTS)0.12[-0.13, 0.38].040.25127
Note.
b represents unstandardized regression weights. SR^2, the squared semi partial correlation.
LL and UL, the lower and upper limits of a confdence interval, respectively. * indicates p < .05.
Table 4 shows the regression model with three variables, two of which were statistically signifcant
with p-values less than 0.05. In addition, the intercept term in the model was also found to be
signifcant. A possible explanation for these efects is as follows: an increase in the proportion
of the uninsured population implies less access to health services, resulting in a decrease in the
overall performance of the health system. The adjusted mortality rate, which is part of the quality
index, had a signifcant negative coefcient, indicating that higher mortality rates were associated
with a decrease in the MDI. This provides an explanation for the low values observed in the Quality
index, given that the calculation of this indicator does not consider the variables
and
.
Finally, the regression model explained 91.5% of the variation in the WDI and, when adjusted,
83% of the variation.
Tests of the statistical assumptions of the regression models were also carried out, including
linearity, normality, homoscedasticity, non-autocorrelation and multicollinearity (Gujarati and
Porter, 2009; Mendenhall et al., 2010). The results are shown in Table 5.
Table 5
Statistical tests
Country
Average
Residues
Durbin-
Watson
Breusch–
Pagan
Shapiro–
Wilk
GVIF
Acceptance
criteria
Colombia-5,118667E-201,930,590,275
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
24
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Country
Average
Residues
Durbin-
Watson
Breusch–
Pagan
Shapiro–
Wilk
GVIF
Costa Rica6,024816E-212,4810,540,31
Dominican
Republic
-3,0116773E-212,310,200,919
Note.
The tests were performed using the R language. The acceptance criteria are established with
a confdence level of 95 %.
Discussion
The results of this research have relevant
implications for policy-making and resource
allocation in the health sector by identifying
areas for improvement in the performance of
the system, such as the low coverage of health
services. The MDI methodology can guide
investment strategies to improve accessibility
and equity in service delivery, as well as in the
infrastructure of the sector, which can have a
positive impact on the population’s access to
primary health care services, not only in terms
of cost reduction, but also in reducing the
distance to nearby health centers, especially in
rural areas, which is often a barrier to accessing
the health system, as highlighted by Bran et al.
(2020).
Based on the above, the research fndings
highlight the importance of improving the
quality of care and delivery of health services
as a determinant of the performance of the
system, to address the shortcomings of the
infrastructure, as well as the availability of
human resources and the rationalization of
process management (Álvarez et al., 2021).
This reinforces the need to optimize the
training of human talent in the health sector
and the availability of resources to ensure
optimal standards of care in order to efectively
reduce mortality, morbidity and malnutrition
rates, thereby improving the quality of life of
the population.
Challenges are also identifed in the
implementation of information and
communication technologies, in the handling
of information, and in the processing of
information for databases in the management
of health data. This represents an opportunity
to improve the efciency of medical care and
the management of clinical information, which
could translate into tangible benefts for all
actors in the system, such as a reduction in
processing time.
However, this must go hand in hand with the
strengthening of public health, so that the
implementation of public programs and policies
can improve the population’s perception of
the quality of health services and strengthen
confdence and legitimacy in the system, thus
encouraging greater community participation.
In this sense, it is vital to emphasize that
the limitations of the study are due to these
challenges, which result in a lack of information
on some variables and therefore hinder the
evaluation of countries such as Cuba and Haiti.
It is clear, consequently, that the study of
health system performance is linked to various
aspects that may vary depending on the space
and time in which they are assessed; it is
therefore essential to approach performance
measurement in a comprehensive manner,
bearing in mind multiple dimensions and
determinants in order to promote signifcant
improvements in the well-being of the
population.
From this point of view, the Multidimensional
Performance Index methodology makes it
possible to identify, from a set of diferent
variables and factors, those that can have the
greatest impact on the health system of the
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
25
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
country, region or city concerned, and also to
compare diferent territorial units on the basis
of the MDI or the dimensions that make it up.
Conclusions
The initial objective was to develop a methodology
that considers the multidimensional nature
of health when measuring the performance
of systems in Latin America. This objective
has been fully achieved through the use of
integrative methods, which show that the
multidimensional measurement methodology
provides relevant information on the behavior
of the health system, which in turn supports
informed decision-making, monitoring of
health indicators and facilitates international
comparisons.
In this sense, the research points to the
need to study the health system from a
multidimensional perspective, with diferent
approaches other than the clinical one.
Likewise, it is necessary to study each country
in its uniqueness, since it can be afected by
diferent factors, as seen in the regression
analyses. These reveal the complexity and
diversity of the variables that infuence the
functioning of these systems, going beyond
the cause-and-efect relationship, for example,
between poverty or resources and health.
However, it should be noted that it is impossible
to capture all of these factors due to the
transversal nature of the object of study and
the stochastic component, although it is
necessary to seek their relationship with other
variables; for example, out-of-pocket spending
on health services or the technological capacity
of clinics and hospitals, while it is true that the
limitation lies precisely in the lack of indicators
and observations for the region.
The importance of the quality of care,
investment in infrastructure, training of human
talent, design and implementation of public
policies, and implementation of information
and communication technologies stand out
as fundamental pillars for strengthening
health systems in Latin America, and it is
therefore imperative that governments and
health institutions continue to work on the
implementation of policies and actions that
address the defciencies identifed by various
measurement tools, including the IDM.
In conclusion, the methodology of the
Multidimensional Index of Health System
Performance for Latin America, although still
in a process of improvement, can help ensure
a health system with equitable access, quality
care and comprehensive coverage, with the
goal of improving the health and well-being of
the population.
Confict of interest
The authors of this article declare that they
have no competing interests that might afect
the reliability of this publication.
References
Agasisti, T., Agostino, D., & Soncin, M. (2020).
Implementing performance measurement
systems in local governments: Moving from
the ‘How’ to the ‘Why’.
Public Performance
& Management Review, 43
(5), 1100-1128.
https://doi.org/10.1080/15309576.2019.170
0806
Álvarez, L., Pérez, M. E., Rodríguez, J. M., &
Palomino, G. P. (2021). La atención al usuario
en los sistemas de Salud [User care in health
systems].
Ciencia Latina, Revista Científca
Multidisciplinar, 5
(2), 1806-1819. https://doi.
org/10.37811/cl_rcm.v5i2.384
Avilés, C. (2017). La pobreza como factor de
riesgo en la salud [Poverty as a health risk
factor]. In R. Zapata, R. Daluuh, V. Cala, & A.
González (Eds.),
Educación, salud y TIC en
contextos multiculturales: nuevos espacios
de intervención
[Education, health and ICT
in multicultural contexts: New spaces for
intervention] (pp. 554–558). Universidad de
Almería. https://dialnet.unirioja.es/servlet/
articulo?codigo=6569963
Behn, R. D. (2003). Why measure
performance? Diferent purposes require
diferent measures.
Public Administration
Review, 63
(5), 586-606. https://doi.
org/10.1111/1540-6210.00322
Bernal, O. & Gutiérrez, C. (comp.).
(2012).
La salud en Colombia: logros,
retos y recomendaciones
[Health in
Colombia: achievements, challenges and
recommendations]
.
Universidad de los
Andes.
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
26
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Bran, L., Valencia, A., Palacios, L., Gómez, S.,
Acevedo, Y., & Arias, C. (2020). Barreras
de acceso del sistema de salud colombiano
en zonas rurales: percepciones de usuarios
del régimen subsidiado [Barriers to access
to the Colombian health system in rural
areas: perceptions of users of the subsidized
regime].
Hacia la Promoción de la Salud,
25
(2), 29-38. https://doi.org/10.17151/
hpsal.2020.25.2.6
Buitrago, G., Amaya, J., Miller, G., & Vera-
Hernández, M. (2023). Cost-sharing in
medical care can increase adult mortality:
Evidence from Colombia.
Center for Global
Development
. https://www.cgdev.org/sites/
default/files/cost-sharing-medical-care-
can-increase-adult-mortality-evidence-
colombia.pdf
Carrillo, V., Jiménez, A., Medina, V., &
González, M. J. (2021). Implementación del
presupuesto público basado en desempeño:
una revisión sistemática de la literatura
empírica [Implementing performance-based
public budgeting: a systematic review of
the empirical literatura].
Revista del CLAD
Reforma y Democracia, 79
, 77-116. https://
doi.org/10.69733/clad.ryd.n79.a223
Comisión Económica para América Latina y
el Caribe (Cepal). (2022). Ficha técnica -
Población en situación de pobreza extrema
y pobreza según área geográfca [Fact sheet
- Population living in extreme poverty and
poverty by geographic área].
CEPALSTAT
Bases de Datos y Publicaciones Estadísticas
.
https://statistics.cepal.org/portal/
cepalstat/technical-sheet.html?indicator_
id=3328&area_id=930&lang=es
De la Guardia, M. A. & Ruvalcaba, J. C. (2020).
La salud y sus determinantes, promoción
de la salud y educación sanitaria [Health
and its determinants, health promotion and
health education].
Journal of Negative and
No Positive Results, 5
(1), 81-90.
Fernández, G., García, M. C., & Escribano, M. C.
(2019). Salud y pobreza, ODS de la Agenda
2030, ¿un reto posible de alcanzar? [Health
and poverty, SDGs of the 2030 Agenda, an
achievable challenge?]
Anales de ASEPUMA
,
(27), 1-29.
Gil de Miguel, Á. & Campuzano, P. (2014).
Pobreza y salud [Poverty and health].
Eu-topías: Revista de Interculturalidad,
Comunicación y Estudios Europeos, 7,
71-78.
https://doi.org/10.7203/eutopias.0.18837
Gujarati, D. N. & Porter, D. C. (2009).
Econometría
[Econometrics] (P. Carril, Trad.;
5
th
ed.). McGraw-Hill Educación.
London, S., Temporelli, K. L., & Monterubbianesi,
P. D. (2009). Vinculación entre salud, ingreso
y educación: Un análisis comparativo para
América Latina [Linking health, income
and education: A comparative analysis
for Latin America].
Economía y Sociedad,
14
(23), 125-146. https://www.redalyc.org/
pdf/510/51011904007.pdf
Marín, Y. A., Higuita, Y., Guerra, D. P., Gómez,
D. A., & Soto, M. L. (2020). Derecho a la salud
en el trabajo: vulneración y fragmentación en
su comprensión y materialización [Right to
health at work: violation and fragmentation
in its understanding and materialization].
Hacia la Promoción de la Salud, 25
(1), 44-59.
https://doi.org/10.17151/hpsal.2020.25.1.4
Mendenhall, W., Beaver, R. J., & Beaver, B.
M. (2010).
Introducción a la probabilidad
y estadística
[Introduction to probability
and statistics] (J. Romo, Trad.; 13
rd
ed.).
CENGAGE Learning.
Ocampo-Rodríguez, M. V., Betancourt, V. F.,
Montoya, J. P., & Bautista, D. C. (2013).
Sistemas y modelos de salud, su incidencia
en las redes integradas de servicios de salud
[Health systems and models, their impact
on integrated health care networks].
Revista
Gerencia y Políticas de Salud, 12
(24), 114–
129.
Peres, F., Blanco, M. P., Monteiro, J., & Brandão,
A. L. (2023). Mapeo de la formación en salud
pública en América Latina: perspectivas
para las instituciones formadoras [Mapping
public health training in Latin America:
perspectives for training institutions].
Revista
Panamericana de Salud Pública, 47
, 1-11.
https://doi.org/10.26633/RPSP.2023.25
Revista Criterios - vol. 31 n.
o
2 Julio-Diciembre 2024 - pp. 12-27
Rev. Criterios ISSN: 0121-8670, e-ISSN: 2256-1161
https://doi.org/10.31948/rev.criterios
Towards a comprehensive measurement of the health system in Latin America:
presentation of the multidimensional performance index
27
Nicolás Rubio García
Jairo Alfredo Ortiz Pacheco
Andrés Mateo Gómez Franco
Rodríguez, R. (2019).
Tratado sobre seguridad
social
[Treaty on social security]
.
Universidad
del Norte.
Serrano, L., Serrano, L., & Rodríguez, A. J. (2019).
Breve análisis conceptual de los términos
salud, enfermedad, prevención y prevención
de enfermedades [Brief conceptual analysis
of the terms health, disease, prevention and
diseases prevention].
Archivos del Hospital
Universitario “General Calixto García”, 7
(3),
439-443.
Sinchi, V. M. (2020). Bioseguridad en el sistema
de salud pública, protección a pacientes
y colaboradores [Biosafety in the public
health system, protection of patients and
collaborators].
Revista Publicando, 7
(25),
39-48.
Vera, C. A., Rodríguez, Y. L., & Hernández,
H. W. (2022). Medición del desempeño del
sistema de gestión de seguridad y salud en
el trabajo: revisión sistemática de literatura
[Measuring occupational safety and
health management system performance:
systematic literature review].
Revista CEA,
8
(18), e2052. https://doi.org/https://doi.
org/10.22430/24223182.2052
Vergara, M. C. (2007). Tres concepciones
históricas del proceso salud-enfermedad.
Hacia la Promoción de la Salud, 12
, 41-50.
World Health Organization (WHO). (2022).
Tracking universal health coverage: 2021
Global Monitoring Report. WHO. https://www.
who.int/publications/i/item/9789240040618
Contribution
Nicolás Rubio García
: Statistical data
processing, writing materials and methods,
and obtaining results. Elaboration of the
methodology. Analysis and interpretation of
results. Contribute to the bibliographic review
and draft the introduction, discussion, and
conclusions.
Jairo Alfredo Ortiz Pacheco
: Review
and analysis of the diferent bibliographic
sources, elaboration of the state of the art
and contribution of the introduction, methods,
discussion and conclusions.
Andrés Mateo Gómez Franco
: Literature
review and contribution to the drafting of the
introduction, discussion, and conclusions.
All authors participated in the preparation of
the manuscript, read it, and approved it.