Articles

Rheumatic heart disease burden, trends, and inequalities in

the Americas, 1990–2017: a population-based study
Pedro Ordunez, Ramon Martinez, Patricia Soliz, Gloria Giraldo, Oscar J Mujica, Porfirio Nordet*

Summary
Background The World Health Assembly 2018 approved a resolution on rheumatic heart disease to strengthen Lancet Glob Health 2019;
programmes in countries where this condition remains a substantial public health problem. We aimed to describe the 7: e1388–97
regional burden, trends, and inequalities of rheumatic heart disease in the Americas. See Comment page e1297
*Dr Nordet retired in August,
Methods In this secondary analysis of the Global Burden of Disease, Injuries, and Risk Factors Study (GBD) 2017, we 2002

extracted data for deaths, prevalence of cases, disability-adjusted life-years (DALYs), years lived with disability, and Department of
Non-Communicable Diseases
years of life lost (YLL) as measures of rheumatic heart disease burden using the GBD Results Tool. We analysed and Mental Health
1990–2017 trends in rheumatic heart disease mortality and prevalence, quantified cross-country inequalities in (P Ordunez PhD, R Martinez Eng,
rheumatic heart disease mortality, and classified countries according to rheumatic heart disease mortality in 2017 and G Giraldo MPH) and
1990–2017. Department of Evidence and
Intelligence for Action in
Health (P Soliz MD,
Findings GBD 2017 estimated that 3 604 800 cases of rheumatic heart disease occurred overall in the Americas in O J Mujica MD), Pan American
2017, with 22 437 deaths. We showed that in 2017 rheumatic heart disease mortality in the Americas was 51% Health Organization,
(95% UI 44–59) lower (1·8 deaths per 100 000 population [95% uncertainty interval 1·7–1·9] vs 3·7 deaths per Washington, DC, USA; and
Cardiovascular Disease Unit,
100 000 population [3·4–3·9]) and prevalence was 30% (29–33) lower (346·4 cases per 100 000 [334·1–359·2] vs Noncommunicable Diseases,
500·6 cases per 100 000 [482·9–519·7]) than the corresponding global estimates. DALYs were half of those globally Noncommunicable Diseases
(55·7 per 100 000 [49·8–63·5] vs 118·7 per 100 000 [108·5 to 130·7]), with a 70% contribution from YLL (39·1 out of and Mental Health, WHO,
55·7 per 100 000). A significant reduction in rheumatic heart disease mortality occurred, from a regional average of Geneva, Switzerland
(P Nordet DSc)
88·4 YLL per 100 000 (95% uncertainty interval 88·2–88·6) in 1990 to 38·2 (38·1–38·4) in 2017, and a significant
Correspondence to:
reduction in income-related inequality, from an excess of 191·7 YLL per 100 000 (68·6–314·8) between the poorest Dr Pedro Ordunez, Department
and richest countries in 1990 to 66·8 YLL per 100 000 (6·4–127·2) in 2017. Of the 37 countries studied, eight (22%) of Non-Communicable Diseases
had both the highest level of premature rheumatic heart disease mortality in 2017 and the smallest reduction in this and Mental Health, Pan
mortality between 1990 and 2017. American Health Organization,
Washington, DC 20037, USA
ordunezp@paho.org
Interpretation The Americas have greatly reduced premature mortality due to rheumatic heart disease since 1990.
These health gains were paired with a substantial reduction in the magnitude of income-related inequalities across
countries, which is consistent with overall socioeconomic and health improvements observed in the Region. Countries
with less favourable rheumatic heart disease situations should be targeted for strengthening of their national
programmes.

Funding None.

Copyright © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND
4.0 license.

Introduction care, resulting in direct and indirect costs to the family

Rheumatic heart disease, a preventable and eradicable
condition1–5 often neglected by media and policy makers, is
a global public health issue.6 33·4 million cases and
319 400 deaths occurred worldwide in 2015.7 The disease
mainly affects vulnerable groups of people living in
certain low-income and middle-income countries,7–9 and
even some groups living in high-income countries.10,11
Rheumatic fever and rheumatic heart disease are the
most common cardiovascular diseases in children,
adolescents, and young adults4,8,12,13 and rheumatic heart
disease has a notable effect on the morbidity and mortality
of pregnant women.8,14 This situation is even more dire
owing to its large economic burden,15,16 which includes the
cost of repeated hospital admissions and valvular surgery
and society.17
The burden of rheumatic heart disease is the
consequence of the synergism of social demographic
disadvantage and poor access to and poor performance of
the health-care system.7,8,10,12,13,18 Prompted by the
Sustainable Development Goals,19 the World Heart
Federation launched a roadmap13 for rheumatic heart
disease prevention and control and the World Health
Assembly approved a resolution6 to develop and imple­
ment national programmes in countries where rheumatic
heart disease remains a substantial public health
problem.
We aimed to produce an assessment of the regional
burden, trends, and inequalities of rheumatic heart

www.thelancet.com/lancetgh Vol 7 October 2019 e1388

 Articles
For the Institute for Health
Metrics and Evaluation gross
domestic product estimates
see http://ghdx.healthdata.org/
record/ihme-data/gross-
domestic-product-gdp-
estimates-country-1950-2015
See Online for appendix
e1389 www.thelancet.com/lancetgh Vol 7 October 2019
Research in context
Evidence before this study
A study from the Global Burden of Disease, Injuries, and Risk
Factors Study (GBD) 2015 reported global, regional, and national
trends in burden of rheumatic heart disease and classified
countries as endemic and non-endemic. However, an assessment
of the rheumatic heart disease burden, trends, and inequalities in
the Americas based on the new estimates from GBD 2017 has
not been done.
Added value of this study
This study quantifies trends in the burden of rheumatic heart
disease in the Americas and 37 of its countries using estimates
from GBD 2017. It also quantifies cross-country inequalities in
rheumatic heart disease burden across the social gradient from
1990–2017 and to identify countries where rheumatic heart
disease remains a public health problem using an
disease in the Americas by use of a comprehensive
approach, which includes a trend analysis of the burden
through a secondary analysis of the Global Burden of
Disease Global Burden of Disease, Injuries, and Risk
Factors Study (GBD) 2017,20 a standard health equity
analytic method favoured by WHO,21 and a classification
of countries based on their pattern of the reduction of
rheumatic heart disease mortality to identify where
rheumatic heart disease remains a public health issue.
Methods
Data sources
We did a secondary analysis of GBD 2017.20 GBD 2017
uses all available up-to-date sources of epi­demiological
data and improved standardised methods to provide a
comparative assessment of health loss across 359 diseases
and injuries by age and sex group for 195 countries and
territories worldwide. GBD uses various interrelated
metrics to measure population health loss, including
number of deaths and mortality, number of cases and
prevalence, years of life lost (YLL) due to premature
death, years lived with disability (YLD), and disability-
adjusted life-years (DALYs). The methods used in the
GBD have been described elsewhere.22–28 For this report,
we extracted estimates and their 95% uncertainty interval
(UI) for deaths, prevalence of cases, DALYs, YLD, and
YLL as measures of rheumatic heart disease (GBD cause
code B.2.1, ICD-10 codes: I01–I01.9, I02.0, I05–I09.9)
burden from GBD 2017 using the GBD Results Tool.28,29
Estimates relevant to rheumatic heart disease from
the GBD 2017 are based on the methods published by
Watkins and colleagues.8 A systematic review was done
for GBD 2013 and updated for GBD 2015; however, no
systematic review was done for 2017. Furthermore, for
GBD 2017 estimation, researchers ran two models using
DisMod-MR: one for non-endemic countries and one
for endemic countries. For GBD 2016, they identified
locations as endemic if the estimated mortality due to
epidemiological classification based on the pattern of reduction
and level of premature mortality.
Implications of all the available evidence
The burden of rheumatic heart disease in the Americas since
1990 has greatly reduced, primarily driven by a reduction in
premature mortality. The observed combination of reductions
in the distributive inequality of the rheumatic heart disease
regional burden, along with the overall reduction in mean
rheumatic heart disease burden, effectively paves the way
towards its elimination as a public health problem. Learning
from countries where rheumatic heart disease control has been
most successful and understanding the social determinants of
rheumatic heart disease inequalities is fundamental to continue
improving regional performance and supporting countries
where challenges in controlling rheumatic heart disease remain.
rheumatic heart disease was greater than 0·15 per
100 000 in the 5–9 years age group, or if that location
had a Socio-demographic Index of less than 0·6. For
GBD 2017, locations were identified as endemic if the
estimated mortality due to rheumatic heart disease was
greater than 0·15 per 100 000 in the 10–14 years age
group or if that location had a Socio-demographic Index
of less than 0·6. In GBD 2016, it was assumed that
there was no remission from rheumatic heart disease.
However, for GBD 2017, remission was estimated in
both the endemic and non-endemic DisMod models.
Such important changes, among others, seem to have
led to an increase in estimated age-specific prevalence
from 33 million cases in GBD 2016 to 40 million cases
in GBD 2017. Detailed methodological information
for rheumatic heart disease estimates and modelling
strategy are published as supplementary material
elsewhere.25
Gross domestic product per capita (in 2005 international
dollars) Institute for Health Metrics and Evaluation
estimates for years 1990 and 2017 were extracted from the
Global Health Data Exchange.29
Data analysis
To characterise the burden of rheumatic heart disease in
the Americas region and the Latin America and the
Caribbean subregion, a descriptive analysis was done.
The number of cases, age-standardised prevalence (per
100 000 population), number of deaths, age-adjusted
mortality (per 100 000 population), and age-standardised
DALYs, YLD, and YLL (per 100 000 population) in both
sexes combined were compared. We also explored
rheumatic heart disease burden across 37 countries and
territories of the Americas. A list of locations is presented
in the appendix (p 6).
We did joinpoint regression analysis30 to assess trends
in rheumatic heart disease burden from 1990 to 2017, by
estimating average annual percentage change (AAPC)
 Articles

for mortality, prevalence, DALYs, YLD, and YLL due to
rheumatic heart disease. We obtained AAPCs and their
95% CI from 1990 to 2017 and subsegments (1990–99,
2000–09, and 2010–17), by age group (5–14 years,
15–49 years, 50–69 years, 70 years and older, and all ages
[age-standardised]), by sex (male, female, and both
sexes combined) at the global, regional, subregional, and
country levels. The AAPC is a summary measure of
the trend over a prespecified fixed interval, which is
computed as a weighted average of annual percentage
change from the joinpoint model, with the weights equal
to the length of the annual percentage change interval.
Monte Carlo permutation method31 with 4499 randomly
permuted data sets was used to obtain the AAPC and its
95% CIs, and the overall asymptotic significance level is
maintained through a Bonferroni correction for a greater
consistency in the p value. These tests also consider
situations with non-constant variance to account for
Poisson variation and autocorrelated errors. AAPC is
considered significant when it is different from zero at
alpha of 0·05. A constant trend was considered when the
zero value was within the 95% CI of the AAPC; an
increasing trend when both limits of the 95% CI were
positive; and a decreasing trend when both limits of the
95% CI were negative.
Cross-country social inequalities analysis
Distributive inequality of rheumatic heart disease burden
across countries was measured by the slope index of
inequality and the health inequality concentration index,
two standard metrics of absolute and relative gradient
inequality, respectively.21 Slope index of inequality was
computed by regressing country-level age-standardised
YLL rates due to rheumatic heart disease in all-ages
population on an income-related relative social position
scale, defined by the midpoint of the cumulative class
interval of the population ranked by gross domestic
product per capita. We used a weighted regression model
to account for heteroskedasticity, and a logarithmic
transformation of the relative social position value to
account for non-linearity due to marginal utility. Health
inequality concentration index was computed by fitting a
Lorenz concentration curve to the observed cumulative
relative distributions of the population ranked by income
and the YLL burden of disease, and numerically
integrating the area under the curve.32

Global The Americas Latin America and the Caribbean

2017
Prevalence
Number of cases 2017 39 345 369 (37 960 759 to 40 828 682) 3 604 800 (3 476 300 to 3 736 570) 3 224 866 (3 104 546 to 3 346 227)
Age-standardised prevalence 2017 (per 500·6 (482·9 to 519·7) 346·4 (334·1 to 359·2) 532·8 (513·2 to 552·8)
100 000 population)
Mortality
Number of deaths 2017 285 517 (266 162 to 303 298) 22 437 (21 702 to 23 251) 7 120 (6 816 to 7 479)
Age-standardised mortality 2017 (per 3·7 (3·4 to 3·9) 1·8 (1·7 to 1·9) 1·2 (1·2 to 1·3)
100 000 population)
Burden of disease
Age-standardised DALYs (per 118·7 (108·5 to 130·7) 55·7 (49·8 to 63·5) 61·9 (52·6 to 73·7)
100 000 population)
Age-standardised years of life lost (per 94·5 (87·5 to 101·4); 80% of DALYs 39·1 (37·8 to 40·8); 70% of DALYs 36·6 (34·8 to 38·8); 59% of DALYs
100 000 population)
Age-standardised years lived with 24·2 (15·7 to 35·2); 20% of DALYs 16·6 (10·8 to 24·1); 30% of DALYs 25·3 (16·4 to 36·8); 41% of DALYs
disability (per 100 000 population)
1990–2017
Prevalence (age-standardised)
Prevalence percentage change 9·7% (9·2 to 10·2) 11·6% (10·8 to 12·5) 0·0% (–0·7 to 0·8)
AAPC from 1990–2017 0·3% (0·3 to 0·3) 0·4% (0·4 to 0·4) 0·0% (0·0 to 0·0)
AAPC from 1990–99 0·5% (0·5 to 0·5) 0·4% (0·4 to 0·4) 0·0% (0·0 to 0·0)
AAPC from 2000–09 0·3% (0·3 to 0·3) 0·4% (0·4 to 0·4) 0·0% (0·0 to 0·0)
AAPC from 2010–17 0·3% (0·2 to 0·3) 0·4% (0·4 to 0·4) 0·0% (0·0 to 0·0)
Mortality (age-standardised)
Total percentage change –54·1% (–58·1 to –51·1) –48·3% (–49·9 to –46·4) –59·0% (–60·8 to –56·7)
AAPC from 1990–2017 –2·9% (–3·0 to –2·8) –2·4% (–2·5 to –2·3) –3·3% (–3·4 to –3·1)
AAPC from 1990–99 –2·8% (–2·9 to –2·7) –2·6% (–2·6 to –2·6) –3·1% (–3·3 to –2·9)
AAPC from 2000–09 –3·3% (–3·6 to –3·1) –2·6% (–2·6 to –2·6) –3·5% (–3·6 to –3·4)
AAPC from 2010–17 –2·4% (–2·5 to –2·3) –1·9% (–2·2 to –1·7) –3·4% (–3·5 to –3·1)
Values are central estimates (95% uncertainty interval). DALYs=disability-adjusted life-years. AAPC=average annual percentage change.

Table 1: Prevalence, mortality, and burden of rheumatic heart disease in 2017 and their trends from 1990 to 2017

www.thelancet.com/lancetgh Vol 7 October 2019 e1390

 Articles

Caribbean Andean Latin Tropical Latin Central Latin Southern Latin High-income Latin America Region of the Global
America America America America North America and Caribbean Americas
800
Cases per 100 000 population

600

Prevalence 400

200 Both
Male
Female
0

10
Deaths per 100 000 population

Mortality 5

400
Cases per 100 000 population

300

Disability-
adjusted 200
life-years

100

40
Per 100 000 population

30

Years lived
with 20
disability
10

400
Per 100 000 population

300

Years of 200
life lost

100

0
90
99
08
17

90
99
08
17

90
99
08
17

90
99
08
17

90
99
08
17

90
99
08
17

90
99
08
17

90
99
08
17

90
99
08
17
20

20

20

20

20

20

20

20

20
19

20

19

20

19

20

19

20

19

20

19

20

19

20

19

20

19

20
19

19

19

19

19

19

19

19

19

Year Year Year Year Year Year Year Year Year

Figure 1: Rheumatic heart disease prevalence, mortality, disability-adjusted life-years, years lived with disability, and years of life lost per 100 000 population by sex at global, regional, and
subregional levels, 1990–2017

Classification of countries based on mortality reduction rates per 100 000 population between 1990–2017 and the
To classify countries by their patterns of reduction in age-standardised YLL rates (per 100 000 population) in
rheumatic heart disease mortality, we did a quadrant 2017 in a population of all ages and both sexes combined.
analysis using the AAPC of the age-standardised YLL We empirically calculated the 33rd and 66th percentiles

e1391 www.thelancet.com/lancetgh Vol 7 October 2019

 Articles

Mortality at 5–14 years of age YLL at all ages

Deaths per 100 000 population Average annual percentage Age-standardised YLL per Average annual percentage
(95% UI), 2017 change (95% CI), 1990–2017 100 000 population (95% UI), change (95% CI), 1990–2017
2017
Global 0·32 (0·29 to 0·36) –4·1% (–4·7 to –3·6) 94·54 (87·50 to 101·38) –3·3% (–3·5 to –3·1)
The Americas Region 0·14 (0·13 to 0·15) –4·2% (–4·3 to –4·0) 39·11 (37·77 to 40·84) –3·1% (–3·2 to –3·0)
Latin America and Caribbean 0·19 (0·17 to 0·20) –4·3% (–4·4 to –4·1) 36·56 (34·81 to 38·79) –3·6% (–3·8 to –3·5)
Antigua and Barbuda 0·15 (0·12 to 0·19) –3·1% (–3·5 to –2·8) 48·92 (44·61 to 54·23) –2·7% (–2·9 to –2·4)
Argentina 0·14 (0·11 to 0·17) –3·1% (–4·1 to –2·2) 87·69 (78·90 to 97·68) –2·3% (–3·1 to –1·5)
Barbados 0·25 (0·19 to 0·33) –2·5% (–4·0 to –1·0) 50·82 (44·85 to 58·15) –2·3% (–2·5 to –2·0)
Belize 0·19 (0·15 to 0·24) –5·1% (–7·3 to –2·7) 61·94 (56·89 to 67·53) –2·5% (–2·8 to –2·3)
Bermuda 0·06 (0·04 to 0·07) –7·4% (–8·1 to –6·6) 18·00 (15·91 to 20·39) –4·9% (–5·0 to –4·7)
Bolivia 0·32 (0·24 to 0·41) –4·7% (–4·8 to –4·6) 72·32 (57·23 to 89·42) –4·6% (–4·7 to –4·5)
Brazil 0·21 (0·19 to 0·23) –4·5% (–4·9 to –4·1) 36·16 (34·82 to 37·76) –3·0% (–3·3 to –2·7)
Canada 0·02 (0·02 to 0·03) –3·6% (–5·3 to –2·0) 22·15 (20·42 to 24·03) –2·6% (–2·8 to –2·5)
Chile 0·03 (0·02 to 0·04) –6·0% (–6·8 to –5·2) 30·77 (27·32 to 34·51) –5·2% (–5·7 to –4·7)
Colombia 0·06 (0·05 to 0·08) –6·3% (–7·8 to –4·9) 11·25 (9·99 to 12·81) –6·2% (–6·9 to –5·4)
Costa Rica 0·10 (0·07 to 0·15) –2·4% (–3·5 to –1·3) 36·21 (32·68 to 41·30) –2·5% (–3·2 to –1·9)
Cuba 0·12 (0·09 to 0·14) –3·8% (–5·3 to –2·3) 46·14 (40·24 to 52·23) –2·6% (–3·3 to –1·8)
Dominica 0·40 (0·31 to 0·52) –1·0% (–1·5 to –0·4) 79·03 (70·31 to 88·45) –2·1% (–2·3 to –1·8)
Dominican Republic 0·42 (0·33 to 0·53) –3·8% (–4·9 to –2·7) 52·72 (45·47 to 60·39) –2·5% (–2·9 to –2·1)
Ecuador 0·28 (0·22 to 0·34) –2·7% (–4·0 to –1·3) 43·52 (38·82 to 48·64) –3·1% (–3·5 to –2·8)
El Salvador 0·18 (0·14 to 0·23) –3·2% (–3·4 to –2·9) 21·52 (18·10 to 25·57) –2·5% (–3·1 to –1·9)
Grenada 0·33 (0·26 to 0·40) –2·8% (–3·2 to –2·4) 104·02 (94·41 to 114·25) –3·4% (–3·6 to –3·2)
Guatemala 0·12 (0·10 to 0·15) –3·9% (–4·4 to –3·3) 18·52 (16·47 to 20·72) –3·1% (–4·2 to –2·0)
Guyana 0·39 (0·31 to 0·48) –3·4% (–4·3 to –2·4) 104·34 (89·19 to 120·61) –2·0% (–2·8 to –1·1)
Haiti 1·31 (1·02 to 1·72) –4·3% (–4·4 to –4·2) 371·21 (290·56 to 476·62) –3·3% (–3·4 to –3·3)
Honduras 0·01 (0·01 to 0·02) –7·2% (–7·7 to –6·7) 9·75 (7·67 to 12·27) –3·1% (–3·3 to –3·0)
Jamaica 0·52 (0·42 to 0·63) –4·9% (–5·3 to –4·5) 62·05 (51·43 to 74·37) –2·8% (–3·8 to –1·8)
Mexico 0·08 (0·07 to 0·08) –4·7% (–5·3 to –4·1) 27·65 (26·41 to 29·18) –4·8% (–5·4 to –4·2)
Nicaragua 0·10 (0·07 to 0·13) –6·6% (–7·1 to –6·1) 17·19 (14·73 to 20·13) –5·1% (–5·4 to –4·7)
Panama 0·13 (0·10 to 0·17) –4·9% (–5·3 to –4·4) 27·09 (24·34 to 29·87) –4·0% (–4·5 to –3·4)
Paraguay 0·14 (0·11 to 0·18) –3·1% (–4·5 to –1·8) 25·34 (21·16 to 30·34) –2·0% (–2·7 to –1·3)
Peru 0·20 (0·15 to 0·25) –3·6% (–5·4 to –1·7) 24·98 (21·05 to 29·10) –3·4% (–4·2 to –2·6)
Puerto Rico 0·04 (0·03 to 0·05) –5·3% (–5·9 to –4·7) 21·34 (19·44 to 23·51) –3·5% (–4·3 to –2·7)
Saint Lucia 0·18 (0·15 to 0·23) –2·7% (–4·0 to –1·4) 74·34 (67·15 to 82·89) –3·4% (–3·9 to –3·0)
Saint Vincent and the Grenadines 0·50 (0·40 to 0·63) 0·0% (–1·0 to 1·0) 79·72 (72·45 to 88·19) –2·1% (–2·7 to –1·4)
Suriname 0·32 (0·26 to 0·40) –3·3% (–4·2 to –2·4) 62·97 (55·46 to 70·69) –2·6% (–2·8 to –2·3)
The Bahamas 0·18 (0·15 to 0·22) –2·7% (–4·1 to –1·3) 58·12 (50·72 to 66·61) –1·9% (–2·1 to –1·6)
Trinidad and Tobago 0·15 (0·12 to 0·19) –5·5% (–7·8 to –3·1) 43·96 (35·17 to 53·20) –3·7% (–4·4 to –3·0)
USA 0·05 (0·04 to 0·05) –1·9% (–2·3 to –1·5) 34·99 (33·37 to 36·62) –2·2% (–2·4 to –1·9)
Uruguay 0·04 (0·03 to 0·04) –3·7% (–4·1 to –3·3) 44·63 (39·74 to 49·78) –2·1% (–2·5 to –1·8)
Venezuela 0·08 (0·06 to 0·10) –4·6% (–5·9 to –3·3) 16·92 (14·56 to 19·75) –4·8% (–5·4 to –4·2)
Virgin Islands 0·09 (0·07 to 0·11) –5·3% (–5·8 to –4·7) 37·39 (32·15 to 43·33) –2·1% (–2·3 to –1·9)
YLL=years of life lost. UI=uncertainty interval.

Table 2: Change in rheumatic heart disease deaths at age 5–14 years and age-standardised YLL at all ages, 1990–2017

(lower and upper terciles) in both measures to classify
countries into nine categories.
Results
GBD 2017 estimated that, in the Americas, 3 604 800 cases
(95% UI 3 476 300 to 3 736 570) of rheumatic heart disease
were estimated to have occurred in the all-age population
in 2017, with 22 437 deaths (21 702 to 23 251) as a result.
Prevalence in the Americas was 30% (29 to 33) lower
(346·4 cases per 100 000 population [95% UI 334·1 to
359·2] vs 500·6 cases per 100  000 population
[482·9 to 519·7]) and mortality was 51% (44 to 59) lower
(1·8 deaths per 100  000 population [1·7 to 1·9] vs
3·7 deaths per 100 000 population [3·4 to 3·9]) than global

www.thelancet.com/lancetgh Vol 7 October 2019 e1392

 Articles

estimates. The age-standardised DALYs in the Americas

A
was equivalent to half of the global level (55·7 per
300 1990
2017 100 000 population [49·8 to 63·5] vs 118·7 per
100 000 population [108·5 to 130·7]), a regional pattern
dominated by premature mortality with 70% (60 to 82)
contribution from YLL (39·1 per 100  000 population
[37·8 to 40·8] of 55·7 per 100 000 population [49·8 to
200 63·5]). Prevalence in the Americas increased from
Years of life lost (per 100 000)

1990 to 2017 at an AAPC of 0·4% (95% UI 0·4 to 0·4),

which is similar to the trend observed globally (0·3%,
95% UI 0·3 to 0·4). In Latin America and the Caribbean
from 1990 to 2017, mortality changed –59·0% (95% UI
–60·8 to –56·7), compared with –48·3% (–49·9 to –46·4)
100 in the Americas and –54·1% (–58·1 to –51·1) globally,
with an AAPC of –3·3% (95% CI –3·4 to –3·1; table 1)
From 1990 to 2017, rheumatic heart disease age-
standardised prevalence and YLD did not change or
slightly increased, a pattern similar to that observed
worldwide (figure 1). Rheumatic heart disease-associated
0
0 0·1 0·2 0·3 0·4 0·5 0·6 0·7 0·8 0·9 1·0 mortality and YLL in the Southern Latin America,
Relative social position by income Caribbean, and Andean Latin America subregions were
R2 1990 0·628 SII 1990 –191·7 (–314·8 to –68·6) higher than in the Americas, although all subregions
R2 2017 0·553 SII 2017 –66·8 (–127·2 to –6·4) showed downward trends between 1990–2017 (figure 1).
B
Both prevalence and mortality were generally higher in
1·0
women than in men but the mortality gap by sex reduced
in all subregions, as had already been observed in the
0·9 high-income countries of North America and Southern
Latin America (figure 1).
0·8 In children aged 5–14 years in 2017, rheumatic heart
Burden of premature mortality (cumulative)

disease-associated mortality in the Americas was

0·7
0·14 deaths per 100  000 population, ranging from
0·6 1·31 deaths per 100 000 population (95% UI 1·02 to 1·72)
in Haiti to 0·02 deaths per 100  000 population
0·5 (0·02 to 0·03) in Canada, and nine countries had a
mortality significantly greater than that of Latin America
0·4
and the Caribbean (0·19 deaths per 100 000 population
0·3 [95% UI 0·17 to 0·20]; table 2). Most countries showed a
significant reduction in mortality from 1990 to 2017,
0·2 with an AAPC for the whole region of –4·2% (95% CI
–4·3 to –4·0), ranging from –7·4% in Bermuda
0·1
(–8·1 to –6·6) to –1·0% in Dominica (–1·5 to –0·4;
0
table 2). However, seven (19%) of 37 countries had an
0 0·1 0·2 0·3 0·4 0·5 0·6 0·7 0·8 0·9 1·0 AAPC significantly smaller than that observed for the
Population share by income (cumulative) entire region (table 2).
HCI 1990 –16·5 (–27·7 to –5·3) The age-standardised YLL at all ages for the region was
HCI 2017 –7·5 (–18·7 to 3·8)
39·11 per 100 000 population (95% UI 37·77–40·84),
C with a wide variation among countries (11·25 per
Equity stratifier Health inequality metrics Year Value 95% CI
Kuznets absolute index (absolute 1990 160·75 159·81 to 161·68
Gross national gap)* 2017 45·55 45·07 to 46·02
product per capita Kuznets relative index (relative gap)* 1990 3·60 3·57 to 3·63 Figure 2: Income-related health inequality regression and concentration
(international $; 2005 constant) 2017 2·36 2·34 to 2·38 curves for the burden of premature mortality due to rheumatic heart disease
Slope index of inequality (absolute 1990 –191·73 –314·82 to –68·64 across the Americas, 1990 and 2017
gradient)* 2017 –66·80 –127·16 to –6·45 (A) Health inequality regression curves. (B) Health inequality concentration
Health concentration index (relative 1990 –16·51 –27·69 to –5·33 curves. (C) Summary measures and 95% CIs for income-related inequalities in
gradient)* 2017 –7·46 –18·69 to 3·77 rheumatic heart disease mortality. SII=slope index of inequality (shown with
Years of life lost to premature mortality 1990 88·43 88·21 to 88·64 95% CI). HCI=health concentration index (shown with 95% CI). *Non-trivially
regional average rate (per 100 000 population) 2017 38·23 38·11 to 38·35† departed from the equity reference. †Statistically significant difference in health
inequality metric estimates between 1990 and 2017.

e1393 www.thelancet.com/lancetgh Vol 7 October 2019

 Articles

Lower tercile: –3·4 Upper tercile: –2·5

110 Low AAPC Medium AAPC High AAPC
Low YLL Low YLL Low YLL
Medium YLL Medium YLL Medium YLL
High YLL High YLL High YLL Grenada Guyana
100

90
Argentina
Saint Vincent and the Grenadines
80
Dominica
Age-standardised YLL per 100 000 population

Saint Lucia
70 Bolivia

Suriname

60 Belize
The Bahamas
Upper tercile: 51·45 Dominican Republic
50 Antigua and Barbados
Barbuda
Median: 40·46 Cuba
Trinidad and Tobago Ecuador Uruguay
40
Virgin Islands
Brazil Costa Rica USA
30 Lower tercile: 27·46
Chile
Mexico Panama Peru Canada Paraguay
20 Bermuda Puerto Rico El Salvador
Guatemala
Nicaragua Venezuela

10 Colombia
Honduras

Median: –2·8
0
–6·5 –6·0 –5·5 –5·0 –4·5 –4·0 –3·5 –3·0 –2·5 –2·0 –1·5 –1·0
Average annual change (%)

Figure 3: Age-standardised YYL per 100 000 in 2017 and average annual percentage change in age-standardised YLL per 100 000 in 1990–2017 due to
rheumatic heart disease for all ages and both sexes in the Americas
Dots represent countries, colour coded according to quadrants defined by lower and upper terciles (33 and 66 percentiles) of the distribution of countries according
to the age-standardised YLL per 100 000 population and the average annual percentage change of the age-standardised YLL rates per 100 000 population in
1990–2017. Haiti, not included in the chart owing to its extreme age-standardised YLL rate (371·21 [95% UI 290·56–476·62]), belongs to Medium AAPC and High YLL
quadrant. AAPC=average annual percentage change. YLL=years of life lost.

100 000 population [9·99–12·81] in Colombia to 371·21 Significant absolute and relative income-related
per 100 000 population [290·56–476·62] in Haiti), and the inequality existed in the burden of premature mortality
YLL for 13 (35%) of 37 countries was significantly greater due to rheumatic heart disease across the 37 countries of
than the region (table 2). Similarly, although all countries the Americas analysed; this burden of premature
showed a clear trend towards YLL reduction between mortality was disproportionately concentrated among
1990 and 2017, ten (27%) of 37 did not achieve significantly poorer countries. A significant reduction in these
lower reductions than those observed for the entire inequalities occurred over time, paired with a reduction
region (table 2). in the regional average from 88·4 YLL per 100 000
Between 1990 and 2017, significant reductions in (95% UI 88·2 to 88·6) in 1990 to 38·2 YLL per 100 000
mortality in each age group occurred, with the greatest (38·1 to 38·4) in 2017. As indicated by the slope index
reduction observed in the 5–14 years age group. Deaths of inequality, an excess of 191·7 YLL per 100  000
and YLL per 100 000 population in 2017, and the AAPC and (314·8 to 68·6) existed between countries with the lowest
95% CI from 1990 to 2017 for mortality and the YLL by age and highest incomes in 1990, which decreased to
group (5–14 years, 15–49 years, 50–69 years, and 70 years 66·8 YLL per 100 000 (127·2 to 6·4) in 2017. Relative
and older) for both sexes combined at the global, regional, gradient inequality, as measured by the health inequality For the Rheumatic Heart
subregional and country levels were computed and are concentration index, was –16·5 (–27·7 to –5·3) in 1990 Disease Interactive Tool see
https://www.paho.org/hq/index.
available in the appendix (p 7). The comprehensive set of and –7·5 (–18·7 to –3·8) in 2017, showing a dis­
php?option=com_content&view
rheumatic heart disease estimates are also available online proportionate concentration of the burden among the =article&id=15246&Itemid=346
in the Rheumatic Heart Disease Interactive Tool. poorer half of the population. The poorest quintile of 5&lang=en

www.thelancet.com/lancetgh Vol 7 October 2019 e1394

 Articles
countries had around 30% of the regional burden of
premature mortality due to rheumatic heart disease in
1990 and 24% in 2017 (figure 2).
12 (32%) of 37 countries, including Haiti (not shown
in the scatter plot because its YLL was above 300 per
100 000), had a YLL in 2017 in the upper tercile of
the distribution (figure 3), which represents the
countries with the highest mortality in 2017. Further­
more, eight of those 12 countries, are in the upper
tercile of the distribution of the AAPC for 1990–2017,
representing the group with the smallest reduction in
mortality: Haiti (highest mortality in the Region),
Guyana, Argentina, Saint Vincent and the Grenadines,
Dominica, Belize, Bahamas, and Dominican Republic.
Suriname also had a small reduction in mortality.
Grenada, Saint Lucia, and Bolivia, although they had a
large YLL in 2017, greatly reduced mortality between
1990–2017. Countries located in the lower tercile of the
distribution, in YLL in 2017 and AAPC 1990–2017
combined, are those with most reduced mortality.
Discussion
Rheumatic heart disease in the Americas has received less
attention in the published literature than other regions of
the world. Our secondary analysis of the GBD 2017 data
offers an updated description of the epidemiology of this
condition in the Americas. The quantification of cross-
country inequalities in rheumatic heart disease burden
across the income gradient adds to the understanding of
its determinants and to the identification of countries
where rheumatic heart disease prevention and control
must be enhanced.
Our study confirms that the Americas is in a better
position than other world regions in terms of rheumatic
heart disease burden, particularly in the reduction
of premature mortality. This trend corresponds with
economic and human development attained by most of its
countries since the late 1990s33 and with the health system
reforms put in place to improve health-care access and
coverage.34,35 The reduction of mortality caused by
rheumatic heart disease is the result of a combination of
factors, such as timely detection, including the extensive
use of echocardiogram as standard, less severe cases, better
treatment, including a wide use of benzathine penicillin,
and greater health-care access and better quality, including
surgical care, among other proximal determinants.
Our finding of a distinctive gradient36 in rheumatic
heart disease burden (ie, the higher the income, the
lower the burden) over time across countries ranked by
their income level was unsurprising, particularly in Latin
America and the Caribbean, one of the most inequitable
regions of the world in terms of income distribution.37,38
However, to the best of our knowledge, such a finding
has not been previously reported in the published
literature. Among all social determinants, wealth or
income is well established as being central to population
health,39,40 and its distributional inequality profoundly
e1395 www.thelancet.com/lancetgh Vol 7 October 2019
affects the distribution of health, affecting access to
health care and health outcomes.41,42
We found a sustained and substantial reduction in
these inequalities along with the reduction of the overall
burden of disease due to rheumatic heart disease; more
specifically, we documented the flattening (or levelling
down) of the slope of YLL due to rheumatic heart disease
mortality in the income hierarchy across countries. This
finding signals the progressivity of the regional reduction:
the closer to the poorest end of the income gradient, the
heavier the rheumatic heart disease burden and the
higher its reduction over time. The diagonalisation (ie,
the tendency of the concentration curve towards the
diagonal line of no inequality or equal distribution) of the
rheumatic heart disease burden inequality concentration
curve over time corroborates this pattern, suggesting that
the gap in rheumatic heart disease burden between
poorer and richer countries is closing. This trend is
consistent with the overall changes promoted by the
Millennium Development Goals in terms of improvement
in access to and coverage of health-care services and in
terms of economic growth, poverty reduction, and social
protection.34,37,38,43
In this regional scenario of success, this study has
identified countries where rheumatic heart disease
remains a public health problem. The level and trend in
premature mortality might become outcome indicators
to evaluate a country’s performance in rheumatic heart
disease control, and they are more robust and stable
when combined. Our study intended to show the
usefulness of this combination, but further analyses are
needed to understand why some countries with similar
economic development, as measured by GDP, have such
different patterns of mortality based on these two
indicators. Countries with vast territories and large
populations will have greater challenges in fighting
rheumatic heart disease, even if their mortality rates are
lower than those of other countries of the region, because
developing and strengthening control programmes,
infrastructure, prevention and education, and treatment
of advanced disease are much more complex and
expensive to implement.
The poor quality and incompleteness of mortality data
for some countries might partially explain why some
countries are in the better performance quadrants of the
scatter plot. However, the use of the GBD methodology20
and its systematically updated estimates28 help in
overcoming this challenge. Additionally, its robust set of
metrics44 (eg, deaths and YLL) are appropriate for
comparisons between countries and regions of the world
and, therefore, for contextualisation of the Americas’
position within the global arena. Nonetheless, the GBD
methodology has some weaknesses. For example,
prevalence estimates are based on a small number of
sources from a few countries and the effect of the
modelling is apparent. Additionally, the joinpoint method
used in this secondary analysis of modelled data rather

than an original analysis of raw data probably leads to
underestimation of the uncertainty in AAPC trends,
because the GBD measure of uncertainty was not
considered in the regression analysis.
The cross-country social inequalities analysis of
rheumatic heart disease burden is not intended for
causal inference. However, it is useful for data
exploration and the extraction of patterns of inequality,45
which might be used to inform policy making. Precisely
because of its exploratory nature, we have deliberately
approached inequality from a bivariate standpoint,
which might also be viewed as a limitation. Formal
confirmatory, multi­variate subnational analysis of social
inequalities in rheumatic heart disease burden is,
therefore, warranted.
The observed combination of non-trivial reductions
in the distributive inequality of the rheumatic heart
disease regional burden and the overall reduction in
mean rheumatic heart disease burden paves the way
towards its elimination as a public health problem.
This highly desirable and achievable goal requires
political will to target the most socially disadvantaged
segments of the population, higher population
awareness and education, sustaining and enhancing
access and coverage, better quality of health services,
including screening, surgery, specialised care, and
valve interventions, and intersectoral action on the
social determinants of health.
Learning from the countries that have been most
successful at controlling rheumatic heart disease and
trying to understand the determinants of differences
among countries is fundamental to continuing to improve
the regional performance and to support countries that
have lagged behind. The full implementation of the World
Health Assembly resolution on rheumatic fever and
rheumatic heart disease is an important opportunity to
revitalise the agenda, consolidate progress, and move
towards the elimination and eradication of rheumatic
heart disease in the Americas.
Contributors
PO and RM developed the original research idea and study design.
PO, RM, and OJM did the data analysis. All authors were involved in
data interpretation and preparation of the manuscript and contributed
important intellectual content. All authors accept accountability for the
overall work.
Declaration of interests
We declare no competing interests.
Data sharing
To download the data used in these analyses, please visit the Rheumatic
Heart Disease Interactive Tool.
Acknowledgments
PO, RM, PS, GG, and OJM are staff members of the Pan American
Health Organization. The authors alone are responsible for the views
expressed in this publication, and they do not necessarily represent the
decisions or policies of the Pan American Health Organization.
References
1 Bach JF, Chalons S, Forier E, et al. 10-year educational programme
aimed at rheumatic fever in two French Caribbean Islands. Lancet
1996; 347: 644–48.
www.thelancet.com/lancetgh Vol 7 October 2019 e1396
Articles
2 WHO. WHO Global Programme for the Prevention of Rheumatic
Fever and Rheumatic Heart Disease. Report of a consultation to
review progress and develop future activities, Geneva,
29 November–1 December 1999. Geneva: World Health Organization,
2000. http://apps.who.int/iris/bitstream/handle/10665/66273/WHO_
CVD_00·1.pdf?sequence=1&isAllowed=y (accessed Nov 26, 2018).
3 Nordet P, Lopez R, Dueñas A, Sarmiento L. Prevention and control
of rheumatic fever and rheumatic heart disease: the Cuban
experience (1986–1996–2002) Cardiovasc J Afr 2008; 19: 135–40.
4 Zühlke L, Karthikeyan G, Engel ME, et al. Clinical outcomes in
3343 children and adults with rheumatic heart disease from
14 low- and middle-income countries: two-year follow-up of the
Global Rheumatic Heart Disease Registry (the REMEDY study).
Circulation 2016; 134: 1456–66.
5 Yusuf S, Narula J, Gamra H. Can we eliminate rheumatic fever and
premature deaths from RHD? Glob Heart 2017; 12: 3–4.
6 WHO. Seventy-First World Health Assembly. Agenda item 12.8.
Rheumatic fever and rheumatic heart disease. Geneva: World
Health Organization, 2018. http://apps.who.int/gb/ebwha/pdf_
files/WHA71/A71_R14-en.pdf (accessed Jan 31, 2019).
7 Watkins DA, Johson CO, Colquhoun S, et al. Global, regional,
and national burden of rheumatic heart disease, 1990–2015.
N Engl J Med 2017; 377: 713–22.
8 WHO. Rheumatic fever and rheumatic heart disease. Report of a
WHO expert consultation, Geneva, 29 October–1 November 2001.
Geneva: World Health Organization, 2004. http://apps.who.int/iris/
bitstream/handle/10665/42898/WHO_TRS_923.pdf (accessed
Nov 26, 2018).
9 Karthikeyan G, Luiza Guilherme L. Acute rheumatic fever. Lancet
2018; 392: 161–74.
10 Jack SJ, Williamson DA, Galloway Y, et al. Prevention of rheumatic
fever in the 21st century: evaluation of a national programme.
Int J Epidemiol 2018; 47: 1585–93.
11 Carapelis J, Brown A, Maguire G, et al. The Australian guideline for
prevention, diagnosis and management of acute rheumatic fever
and rheumatic heart disease (2nd edn). Casuarina: Menzies School
of Health Research, 2012. https://www.rhdaustralia.org.au/
node/950/attachment (accessed Jan 31, 2019).
12 Remenyi B, Carapetis J, Wyber R, Taubert K, Mayosi BM,
World Heart Federation. Position statement of the World Heart
Federation on the prevention and control of rheumatic heart
disease. Nat Rev Cardiol 2013; 10: 284–92.
13 Palafox B, Mocumbi AO, Kumar RK, et al. The WHF roadmap for
reducing CV morbidity and mortality through prevention and
control of RHD. Glob Heart 2017; 12: 47–62.
14 Mocumbi AO, Jamal KK, Mbakwem A, Shung-King M, Sliwa K.
The Pan-African Society of Cardiology position paper on
reproductive healthcare for women with rheumatic heart disease.
Cardiovasc J Afr 2018; 29: 394–03.
15 Milne RJ, Lennon D, Stewart JM, Vander Hoorn S, Scuffham PA.
Mortality and hospitalisation costs of rheumatic fever and
rheumatic heart disease in New Zealand. J Paediatr Child Health
2012; 48: 692–97.
16 Soudarssanane MB, Karthigeyan M, Mahalakshmy T, et al.
Rheumatic fever and rheumatic heart disease: primary prevention
is the cost effective option. Indian J Pediatr 2007; 74: 567–70.
17 Robertson KA, Mayosi BM. Rheumatic heart disease: social and
economic dimensions. S Afr Med J 2008; 98: 780–81.
18 Dougherty S, Beaton A, Nascimento BR, Zühlke LJ, Khorsandi M,
Wilson N. Prevention and control of rheumatic heart disease:
overcoming core challenges in resource-poor environments.
Ann Pediatr Cardiol 2018; 11: 68–78.
19 UN. Sustainable Development Goals. Sustainable Development
Goals Knowledge Platform, 2017. https://sustainabledevelopment.
un.org/sdgs (accessed Dec 20, 2018).
20 Institute of Health Metrics and Evaluation. Global Burden of
Disease Study 2017 (GBD 2017) data resources. Institute of Health
Metrics and Evaluation, 2017. http://ghdx.healthdata.org/gbd-2017
(accessed Nov 8, 2018).
21 WHO. Handbook on health inequality monitoring, with a special
focus on low- and middle-income countries. Geneva: World Health
Organization, 2013.
22 Murray CJL, Ezzati M, Flaxman AD, et al. GBD 2010: design,
definitions, and metrics. Lancet 2012; 380: 2063–66.
 Articles
23 GBD 2017 Causes of Death Collaborators. Global, regional,
and national age-sex-specific mortality for 282 causes of death in
195 countries and territories, 1980–2017: a systematic analysis for
the Global Burden of Disease Study 2017. Lancet 2018; 392: 1736–88.
24 GBD 2017 Risk Factors Collaborators. Global, regional, and national
comparative risk assessment of 84 behavioural, environmental and
occupational, and metabolic risks or clusters of risks for
195 countries and territories, 1990–2017: a systematic analysis for
the Global Burden of Disease Study 2017. Lancet 2018; 392: 1923–94.
25 GBD 2017 Disease and Injury Incidence and Prevalence
Collaborators. Global, regional, and national incidence, prevalence,
and years lived with disability for 354 diseases and injuries for
195 countries and territories, 1990–2017: a systematic analysis for the
Global Burden of Disease Study 2017. Lancet 2017; 392: 1789–858.
26 GBD 2016 Mortality Collaborators. Global, regional, and national
under-5 mortality, adult mortality, age-specific mortality, and life
expectancy, 1970–2016: a systematic analysis for the Global Burden
of Disease Study 2016. Lancet 2017; 390: 1084–150.
27 GBD 2017 DALYs and HALE Collaborators. Global, regional, and
national disability-adjusted life-years (DALYs) for 359 diseases
and injuries and healthy life expectancy (HALE) for 195 countries
and territories, 1990–2017: a systematic analysis for the Global
Burden of Disease Study 2017. Lancet 2018; 392: 1859–922.
28 Global Burden of Disease Collaborative Network. GBD results tool.
Seattle: Institute for Health Metrics and Evaluation, 2017.
http://ghdx.healthdata.org/gbd-results-tool (accessed Nov 20, 2018).
29 James SL, Gubbins P, Murray CJL, Gakidou E. Developing a
comprehensive time series of GDP per capita for 210 countries
from 1950 to 2015. Popul Health Metr 2012; 10: 12.
30 National Cancer Institute. Joinpoint Trend Analysis Software
version 4·6.0·0. National Cancer Institute, 2018.
https://surveillance.cancer.gov/joinpoint/ (accessed Jan 16, 2019).
31 Kim HJ, Fay MP, Feuer EJ, Midthune DN. Permutation tests for
joinpoint regression with applications to cancer rates. Stat Med
2000; 19: 335–51.
32 Mujica OJ, Moreno CM. From words to action: measuring health
inequalities to “leave no one behind”. Pan Am J Public Health 2019;
43: e12.
e1397 www.thelancet.com/lancetgh Vol 7 October 2019
33 The World Bank. World Bank country and lending groups for the
2019 fiscal year. The World Bank, 2018. https://datahelpdesk.
worldbank.org/knowledgebase/articles/906519-world-bank-country-
and-lending-groups (accessed Oct 22, 2018).
34 Atun R, de Andrade LO, Almeida G, et al. Health-system reform and
universal health coverage in Latin America. Lancet 2015; 385: 1230–47.
35 Frenk J, Gómez-Dantés O, Knaul FM. The democratization of
health in Mexico: financial innovations for universal coverage.
Bull World Health Organ 2009; 87: 542–48.
36 Marmot M. The health gap: the challenge of an unequal world.
Lancet 2015; 386: 2442–44.
37 Pan American Health Organization. Health in the Americas+, 2017
edition. Summary: regional outlook and country profiles.
Washington, DC: Pan American Health Organization, 2017.
38 Economic Commission for Latin America and the Caribbean. Social
Panorama of Latin America 2018. Santiago: Economic Commission
for Latin America and the Caribbean, 2019.
39 The World Bank. Poverty and shared prosperity 2016: taking on
inequality. Key findings. Washington, DC: The World Bank, 2016.
40 Doyle MW, Stiglitz JE. Eliminating extreme inequality: a sustainable
development goal, 2015–2030. Ethics Int Affairs 2014; 28: 5–13.
41 Wilkinson RG, Pickett KE. Income inequality and population
health: a review and explanation of the evidence. Soc Sci Med 2006;
62: 1768–84.
42 Subramanian SV, Kawachi I. Income inequality and health:
what have we learned so far? Epidemiol Rev 2004; 26: 78–91.
43 Marmot M, Filho AP, Solar O, Fortune K. Action on the social
determinants of health in the Americas. Pan Am J Public Health
2013; 34: 379–81
44 Martinez R, Soliz P, Caixeta R, Ordunez P. Reflection on modern
methods: years of life lost due to premature mortality—a versatile
and comprehensive measure for monitoring non-communicable
disease mortality. Int J Epidemiol 2019; published online Jan 9.
DOI:10.1093/ije/dyy254.
45 Tukey JW. Exploratory data analysis. Reading, PA: Addison-Wesley,
1977.