NOTHING ON RECOVERY! Obviously they don't give a flying fuck about stroke survivors.
Since you didn't measure 100% recovery I
guess that recovery is not important to you. Maybe you want to talk to
survivors sometime, they'll give you an earful.
Measure recovery and results, NOT cases of stroke. I'd fire everyone involved in this.
“What's measured, improves.” So said management legend and author Peter F. Drucker
The latest crapola here:
Global stroke statistics 2022
Abstract
Background:
Contemporary
data on stroke epidemiology and the availability of national stroke
clinical registries are important for providing evidence to improve
practice and support policy decisions.
Aims:
To
update the most current incidence, case-fatality, and mortality rates
on stroke and identify national stroke clinical registries worldwide.
Methods:
We
searched multiple databases (based on our existing search strategy) to
identify new original papers, published between 1 November 2018 and 15
December 2021, that met ideal criteria for data on stroke incidence and
case-fatality, and added these to the studies reported in our last
review. To identify national stroke clinical registries, we updated our
last search, using PubMed, from 6 February 2015 until 6 January 2022. We
also screened reference lists of review papers, citation history of
papers, and the gray literature. Mortality codes for International
Classification of Diseases (ICD)-9 and ICD-10 were extracted from the
World Health Organization (WHO) for each country providing these data.
Population denominators were obtained from the United Nations (UN) or
WHO (when data were unavailable in the UN database). Crude and adjusted
stroke mortality rates were calculated using country-specific population
denominators, and the most recent years of mortality data available for
each country.
Results:
Since
our last report in 2020, there were two countries (Chile and France)
with new incidence studies meeting criteria for ideal population-based
studies. New data on case-fatality were found for Chile and Kenya. The
most current mortality data were available for the year 2014 (1
country), 2015 (2 countries), 2016 (11 countries), 2017 (10 countries),
2018 (19 countries), 2019 (36 countries), and 2020 (29 countries). Four
countries (Libya, Solomon Islands, United Arab Emirates, and Lebanon)
reported mortality data for the first time. Since our last report on
registries in 2017, we identified seven more national stroke clinical
registries, predominantly in high-income countries. These newly
identified registries yielded limited information.
Conclusions:
Up-to-date
data on stroke incidence, case-fatality, and mortality continue to
provide evidence of disparities and the scale of burden in low- and
middle-income countries. Although more national stroke clinical
registries were identified, information from these newly identified
registries was limited. Highlighting data scarcity or even where a
country is ranked might help facilitate more research or greater policy
attention in this field.
Introduction
The
worldwide burden of stroke remains massive, and there is a continued
need to understand trends of the disease and its impact on each country,
to guide policy decisions and healthcare planning. We have previously
reported stroke statistics by country.1–3
We have also previously recognized the value of national stroke
clinical registries, not just for routine monitoring of patient
characteristics, access to clinical care, and health outcomes, but also
their potential to be a reliable supplement or substitute for
epidemiological studies in countries where a large proportion of the
population is hospitalized.3,4
This fourth issue is part of an ongoing series to update information on
(1) stroke incidence, case-fatality and mortality, and availability of
national stroke clinical registries around the world; and (2) highlight
where data are lacking or outdated.
Aims
Our
aims were to (1) update our repository of the most recent
country-specific data on stroke; (2) determine where stroke incidence,
case-fatality, and mortality are new, old, or missing; and (3) determine
any newly identified national stroke clinical registries.
Methods
We
used similar methods to those described in our previous reviews to
update stroke incidence, case-fatality, mortality, and national stroke
clinical registries.1–3 These methods are described below. A summary of all data sources is provided in Supplementary Table 1.
Literature search and data extraction for stroke incidence and case-fatality
For
stroke incidence and case-fatality, we searched Medline, Scopus,
PubMed, and Google Scholar to identify relevant original manuscripts and
reviews for articles between 1 November 2018 and 15 December 2021. We
included studies with data on overall stroke incidence, incidence among
men and women separately, or 28- or 30-day case-fatality. Consistent
with our previous reviews, citations identified were first screened by
title and abstract (T.T.). After this, full texts were read by a single
reviewer (T.T.). We included only original manuscripts meeting the ideal
criteria (Supplementary Table 2).5
We also screened the reference lists of included manuscripts and
relevant review articles, and discussed identification of any
potentially missing original manuscripts that could have data on stroke
incidence or case-fatality and would meet the inclusion criteria. For
the heat maps on incidence, when there were data from the same country
from rural and urban regions, we present the mean value of these
observations for the country. Information from each manuscript was
extracted by one reviewer (T.T.). All authors discussed the list of
identified manuscripts.
Data extraction for stroke mortality
All mortality data were obtained from the World Health Organization (WHO) website (https://www.who.int/data/data-collection-tools/who-mortality-database).
These data are provided to the WHO by each country, having arisen from
in-country civil registration systems and coded by their national
authorities. We used the latest available data files on country codes
and mortality from the WHO, these being updated on 15 March 2022 and
downloaded on 27 May 2022. We used mortality coded using the
International Classification of Diseases (ICD) versions 9 and 10 (ICD-9
and ICD-10; Supplementary Table 3).
Population
data (i.e. denominators) for the year corresponding to the latest
mortality data were used to calculate mortality rates. In contrast to
previous reviews,1–3 population data were obtained largely from the United Nations (UN) database.6
UN population data were preferred to the WHO population data because of
the lack of up-to-date corresponding population data in the WHO
database. When population data for the corresponding mortality year were
unavailable in the UN database, population data from the WHO were
substituted. For countries in which mortality data were reported to the
WHO by specific subdivisions of the country, for example, China rural
and urban areas, we used the WHO population data for the corresponding
subdivision and mortality year.
Literature search and data extraction for national stroke clinical registries
For
the identification of national stroke clinical registries around the
world, we undertook a comprehensive search of the peer-reviewed
literature in PubMed. Registries that were considered to represent a
national standardized dataset for acute stroke care and outcomes in
their country were included. Our search was limited to information
available in English language. We leveraged results of the previous
literature search4 and undertook an updated search from 6 February 2015 to 6 January 2022. A detailed description of methods is provided in online supplementary methods.
Analysis of data
We compared stroke incidence with the proportion of the population aged ⩾65 years, as previously.1
For each country, we used the UN population data to determine the
proportion aged ⩾65 years from the same year as the incidence data. When
incidence was assessed over more than 1 year, we used the population
data corresponding to the mid-year of data collection. When an incidence
study spanned an even number of years, the more recent of the
2 mid-years were used. However, when there were no population data for
the year in which the incidence study was conducted, we used the closest
available year, up to 2 years. A regression analysis was undertaken to
determine the association between the proportion of the population aged
⩾65 years and the overall crude incidence, and a p value <0.05
was considered statistically significant. We provide a new analysis
comparing case-fatality (death within 28–30 days post-stroke) by the
proportion of the population aged ⩾65 years, using a similar approach to
analyses of incidence.
For stroke mortality,
we calculated overall and sex-specific rates (100,000 population/year),
using previously described methods.1–3
Crude mortality rates were also compared by the proportion of the
population aged ⩾65 years, using a similar approach to analyses of
incidence and case-fatality.
Death rates were age-standardized to the “new world” population,7
using the direct method and 5- or 10-year age bands, depending on how
population data were reported by country. Countries varied in their
report of upper age bands for mortality, ranging from ⩾65 years to
⩾95 years. In each instance, we adjusted for age using the maximum
number of categories available for each country. A few countries (South
Africa, Tunisia, Argentina, Bolivia, Brazil, Mexico, Saint Lucia)
reported mortality data for unknown sex. In such instances, we allocated
cases proportionally to male or female, depending on the distribution
of deaths in each sex and age categories. Age-adjusted death rates were
used to compare mortality between countries/regions and to estimate
percentage change in death rate/year between current and last reviews.
To ensure reliable comparison of rates between the current and last
review, population denominators corresponding to the years under review
were obtained from the same source. Furthermore, analysis of change in
death rates was stratified by population size (based on a population of
less than vs at least 5 million), as change estimates may not be robust
when based on small population denominators.
Consistent with our previous reviews,1–3
a regression analysis was undertaken to assess the association between
the crude incidence rate and the corresponding crude mortality rate for
each country, using the same year for incidence and mortality. When the
study was conducted over more than 1 year, we used the most relevant
population data, as described above for analyses of incidence and
case-fatality. Data on national stroke clinical registries were
summarized using descriptive statistics.
Results
Overall,
since our last review, updated incidence data that fully or partially
met the inclusion criteria were identified. New case-fatality data were
also found. We were able to source new mortality data for several
countries and for more recent time periods (as described below). Several
national stroke clinical registries were also newly identified.
New literature on stroke incidence
The initial literature search returned 330 manuscripts (Figure 1 and Supplementary Tables 4–7).
After the first stage of screening, we included 58 potential new
manuscripts on stroke incidence for full text review. Since our last
review,1
two new studies that fully met the inclusion criteria were identified
(France and Chile), and new incidence data that partially met the
inclusion criteria were available for one country (China). Figure 2
is the heat map showing adjusted incidence rates of stroke worldwide.
Countries having old, new, and no data for overall stroke incidence and
sex-specific incidence are illustrated (Supplementary Figures 1 and 2).
In Chile, the overall crude stroke incidence rate in Ñuble (2015–2016)
was 180/100,000/year, being greater for men (186/100,000/year) than
women (175/100,000/year) (Supplementary Tables 4 and 5).
The overall adjusted incidence rate for Chile (Ñuble) was
122/100,000/year. These figures are considerably larger than those
obtained in Iquique in 2000–2002 (crude incidence 74/100,000/year and
adjusted incidence 86/100,000/year). In an updated study conducted in
2002–2012 in Dijon, France, the overall adjusted incidence was
116/100,000/year, 126/100,000/year in men and 78/100,000/year in women.
These figures were slightly greater than age-adjusted rates from 2000 to
2006 in the same region.
The
study from Tianjin, China, only partially met the inclusion criteria
(i.e. conducted among people aged ⩾65 years), and therefore, we could
not reliably compare incidence estimates to those of other countries.
Nevertheless, the sex-specific age-adjusted incidence among those aged
⩾65 years was greater in men (2875/100,000/year) than women
(1839/100,000/year; Supplementary Tables 6 and 7).
Consistent
with our previous report, there was a strong linear relationship
between crude incidence rates and the percentage of the population aged
⩾65 years in both low- and middle-income countries (LMICs) and
high-income countries (HICs; Figure 3(a)).
In regression analyses undertaken between crude incidence rate and the
study year, there was a negative relationship in HICs, but a strong
positive relationship for LMICs (Figure 3(b)).
Updated data on age-adjusted stroke incidence are primarily from HICs
and continue to be less than age-adjusted rates observed in LMICs (Supplementary Figure 3(a)–(c)).
New literature on stroke case-fatality
Our literature search returned 329 manuscripts (Figure 1). Only two new manuscripts with case-fatality data met our inclusion criteria (Supplementary Tables 8 and 9). In Kenya (Nairobi), the overall case-fatality at 28 days was 26.7%,8 while that reported at 30 days in Chile (Ñuble) was 24.6%.9 There continues to be a significant variation in stroke case-fatality both between and among LMICs and HICs (Figure 4), with case-fatality being greater in men than women in some countries, and greater in women than men in other countries (Figure 5).
In a new analysis, we identified that people aged ⩾65 years had reduced
case-fatality rates, although this association was not significant when
stratified into LMICs and HICs (Figure 6(a)). No significant association was found between stroke case-fatality and the year the studies were undertaken (Figure 6(b)).
Updated mortality statistics from around the world
Mortality data are reported for 138 countries (Supplementary Tables 10 and 11).
Since our last review, where the most recent data available were for
2017, there are now 108 countries for which new data are available,
including four countries (Libya, Solomon Islands, United Arab Emirates,
and Lebanon) for which mortality data were available for the first time.
There are updated data for 2014 (Sri Lanka), 2015 (Bahamas, Saint
Lucia), 2016 (11 countries), 2017 (10 countries), 2018 (19 countries),
2019 (36 countries), and 2020 (29 countries). There are 30 countries for
which there were no new data available, with the oldest data (1983)
being from the Falkland Islands (Malvinas; Figure 7).
There were 11 countries which reported mortality data using a broad
category of “cerebrovascular disease,” including conditions other than
stroke (Supplementary Table 11).
For
estimating mortality rates, we used population data (denominators)
obtained from the UN database for 129 countries and population data from
the WHO database for 9 countries (Supplementary Table 10). This was in contrast to the last review,1
in which WHO population denominators were used for 123 countries and UN
population denominators for 15 countries. Moreover, in contrast to the
last review,1
Pakistan was excluded in the current analysis due to lack of
corresponding population data. While crude mortality estimates were
largely unaffected by the change in population data (i.e. from WHO data
in the 2019 review to the UN data in the current review), there were
considerable changes in age-adjusted mortality estimates (by
⩾5/100,000/year) for some countries (Supplementary Table 12).
Countries with the largest crude mortality rates were mostly from Eastern or South-Eastern Europe (Figure 7).
Excluding the Falkland Islands which had old data (1983), the top 14
countries with the largest crude mortality rates were from these
European regions, all of which had fairly recent data (2018–2020).
Similar to our last review,1 Bulgaria (294 deaths/100,000/year; 2019) and Latvia (271/100,000/year; 2020) have the greatest crude stroke mortality rates (Supplementary Table 11).
Georgia moved seven places up to the third position (254/100,000/year;
2020), overtaking Romania which now sits fifth on the list
(199/100,000/year; 2019). By contrast, there was a considerable decline
for Serbia from 2015 to 2020, dropping 11 places to sit in the 19th
position (Figure 7).
In contrast, with the exception of Solomon Islands (2018), countries
with newer data at the lowest end of the spectrum (<15/100,000/year)
were from the Middle-Eastern region, including the United Arab Emirates
(2020), Qatar (2020), Oman (2019), and Morocco (2016). Papua New Guinea
and Haiti had the smallest mortality rates, but data were more than
17 years old.
There was a significant
association between the proportion of the population aged ⩾65 years and
crude mortality rate (slope = 4.05, 95% confidence interval (CI) =
2.84–5.26; p < 0.001; Figure 8). However, this relationship was steeper in LMICs (slope = 8.90; 95% CI = 6.81–10.99; p < 0.001) than in HICs (slope = 3.95, 95% CI = 2.50–5.40; p < 0.001).
Andorra (2020), Iceland (2020), Canada (2019), Australia (2020), and
Puerto Rico (2017) had very low crude mortality rates
(<40/100,000/year), despite having ⩾15% of their populations aged
⩾65 years. Countries with the greatest mortality rates
(⩾100/100,000/year), despite having <15% of the population aged
⩾65 years, included the Republic of North Macedonia (2020), Belarus
(2018), the Republic of Moldova (2018), China (all three subdivisions;
2000), and Azerbaijan (2007). There was no significant linear
relationship between the year mortality data were collected and crude
mortality rates for LMICs or HICs (data not provided).
When adjusting mortality rates to the new world population (Figures 9 and 10),
countries with new data available and adjusted mortality rates of
stroke ⩾100/100,000/year included Bulgaria (2019) and Kyrgyzstan (2019),
TFYR Macedonia (2020), Republic of Moldova (2018), and Russian
Federation (2019). Overall, 78/108 (72%) countries with new data had a
reduced adjusted mortality rate since our last review.1
The largest changes in death rates were observed in small countries, that is, with <500,000 population (Supplementary Table 13).
These included a decline in rate by ⩾10%/year for Antigua and Barbuda
(2016–2017), Seychelles (2015–2019), and Malta (2015–2017) and an
increase by ⩾20%/year for Virgin Islands (USA; 2016–2017), Martinique
(2015–2016), Reunion (2015–2016), Saint Vincent and Grenadines
(2016–2017), and Guadeloupe (2015–2016). Among countries with ⩾5 million
population (Supplementary Table 14),
Egypt (2015–2019), Morocco (2014–2016), Brazil (2016–2019), Dominican
Republic (2013–2018), Turkey (2019), and Republic of Korea (2016–2019)
led the group, with declines in adjusted death rates by ⩾6.5%/year. In
contrast, Tajikistan (2016–2017), Peru (2015–2018), Iran (Islamic
Republic; 2015–2016), and Venezuela (2013–2016) were countries with ⩾5
million population and large increase in adjusted rates, by ⩾5%/year.
There
were 65 incidence studies from 37 countries in which overall incidence
and corresponding mortality data were available. There was a strong
positive relationship between incidence and mortality rates from stroke
(slope = 1.22, 95% CI = 0.46–1.19; p < 0.001), and this association was similar for HICs and LMICs (Figure 11).
Update on national stroke clinical registries
The
initial search returned 2659 manuscripts. After removing duplicates,
2656 studies were screened for inclusion, of which 362 publications met
the eligibility criteria (Supplementary Figure 4).
Most (93%) of the 362 publications related to national stroke clinical
registries were for updated data and/or analyses of data from the 28
national stroke clinical registries that were previously identified.
From 25 publications, there were seven national stroke clinical
registries, predominantly from HICs, that were newly identified. There
was limited information from the newly identified national stroke
clinical registries. Moreover, no case report forms were published, and
relevant information was lacking in associated websites and other
references (Supplementary Tables 15 and 16, Supplementary Figures 1 and 4).
Discussion
In
this comprehensive updated review of global stroke statistics, we
present the most recently available country-specific stroke incidence,
case-fatality, mortality data, and newly identified national stroke
clinical registries. For the first time, we determined the relationship
of crude case-fatality with the percentage of the population aged
⩾65 years and the year the studies were conducted.
Compared with our last review,1
updated incidence data were only available in two countries, Chile and
France, both HICs. The scarcity of data on stroke incidence in other
countries seems to be consistent with previous findings.1,3 A shift in the burden of stroke from HICs to LMICs was previously highlighted.1
In this review, we observed that crude incidence continues to decrease
over time in HICs and increase over time in LMICs, which is a cause for
concern. In addition, there continues to be an overall variation in
estimates of stroke incidence across different regions of the world.
High-quality data are needed to understand stroke burden, and to help
plan and develop country-level strategies to improve stroke care.10
The need for well-designed, community-based stroke surveillance studies
meeting the ideal criteria is equally important in HICs and LMICs.
Data
on estimated overall crude case-fatality rates were available for 23
countries. Of interest is that case-fatality seems to be increasing in
LMICs, compared with HICs, highlighting the urgency to improve stroke
care in LMICs. Overall, there continues to be variation in case-fatality
across different regions. Recommendations to provide crude and
age-adjusted case-fatality estimates for different age and sex
categories would ensure a better understanding of trends in
stroke-specific deaths.
In contrast to our last review,1
we identified more recent mortality data for several countries.
However, countries with the greatest crude mortality rates remain
largely unchanged. Similarly, countries with the largest age-adjusted
mortality rates remain unchanged, mainly because no new data were
available for many countries. There were considerable declines in
age-adjusted mortality rates over time for several countries, notably
Egypt, Serbia, Morocco, Brazil, Dominican Republic, Turkey, and Republic
of Korea. In contrast, substantial increases in age-adjusted rates were
observed for Tajikistan, Peru, Iran, and Venezuela. We observed a
geographical pattern in crude mortality rates at both ends of the
spectrum. While countries with the highest crude mortality rates were
largely from Eastern or South-eastern Europe, those with the lowest
rates were largely from the Middle-Eastern region. For countries with
reduced mortality rates, such as Haiti and Papua New Guinea, estimates
may be partly explained by the fact that data are old, so do not reflect
any recent changes that may have occurred. Compared with our last
review,1
the association between the proportion of the population aged ⩾65 years
and crude mortality rate slightly decreased, and this association
remained stronger in LMICs than HICs.
Countries
with the greatest mortality rates are largely unchanged, suggesting a
need for more effective policy actions for primary and secondary
prevention in these countries. The decline or stability in crude
mortality rates over time for most countries, particularly in middle-
and high-income countries, is likely to be attributed to declines in
stroke incidence11
and improved strategies for the management of stroke (e.g. acute stroke
units, better identification of milder strokes). We also observed
greater crude mortality rates among countries in Eastern and
South-Eastern Europe, despite having relatively younger population. This
finding may be partly explained by the relatively low national
income/person, when compared with other countries in which mortality
data were available.12
In contrast, the relatively low crude mortality rates in the more
developed/advanced countries in the Middle-eastern region, compared with
other developed/advanced countries, may be due to a combination of
structural factors that enhance primary and secondary prevention.13
These may include factors, such as improved socioeconomic conditions
and quality of care, advocacies and favorable changes in policies and
legislations, and healthy and safer environment, which may reduce the
incidence of stroke or any associated death.10
Some
countries with considerable mortality rates, for example, the Russian
Federation, Falkland Islands (Malvinas), Ukraine, and China, used a
broad category of “cerebrovascular disease,” including conditions other
than stroke. Therefore, these rates may be overestimated. Such countries
should be encouraged to report mortality using stroke-specific codes to
enable better comparability between countries. We observed substantial
discrepancies between population data obtained from the UN and that from
the WHO, for a few countries that were analyzed using the same
mortality data, but had population denominators from different sources.
Notably, South Africa, Azerbaijan, the Syrian Arab Republic, and small
Caribbean and African countries had different population denominator
data. The effect of these discrepancies on adjusted mortality estimates
highlights the need for countries to report high-quality and more
complete data, regardless of database, to reliably inform relevant
domestic and global policy decisions and actions.
National
stroke clinical registries are becoming more common with seven newly
identified registries since the last systematic search. However, most
were from HICs, with five being from Europe. Although more national
stroke clinical registries were identified, information from these newly
identified registries was limited. Importantly, national stroke
clinical registries were located largely in countries with either no
studies on incidence or with no new studies on stroke incidence, thereby
expanding the number of countries with some type of up-to-date stroke
data.
There may be several reasons why there
is underrepresentation of national stroke clinical registries in certain
regions. First, monitoring the quality of stroke care provided in
hospitals may not be a priority in some countries if they do not have
the resources or access to a system for data collection. Second, there
may be a preference to conduct resource-intensive community-based stroke
incidence studies within a geographical region to capture events in and
out of hospitals to quantify the burden of stroke. Third, our criteria
to consider a stroke registry to be “national,” while flexible, was not
intended to capture regional registries within a country, even though
some of these registries may have had better coverage than some
considered to be “national.”
Finally, the
Registry of Stroke Care Quality (RES-Q) is gaining prominence as a
global tool for continuous monitoring, evaluation, and improvement of
healthcare quality.14
RES-Q comprises data on 83 countries (~500,000 patients, 1838 acute
hospital sites), including LMICs, and is an example of an international
resource supporting standardized collection of stroke data. Importantly,
this tool reduces costs of establishing the infrastructure for national
clinical quality registries. However, site- and national-level coverage
is unclear.
The main strengths of this
review include its comprehensiveness and the use of high-quality studies
that met strict criteria. Our review is limited by the fact that stroke
incidence and case-fatality data are often available only in certain
regions of a country and may not be representative of the whole country.
Because mortality rates have been declining over time, the rank
position of a country may also be influenced by the availability of
recent population data, differences in data collection policies for
reporting deaths, and potential misclassification of causes of death.
Conclusion
Up-to-date
information on stroke incidence, case-fatality, and mortality continues
to provide evidence of disparities among countries and changing
magnitudes of burden, particularly in LMICs. Although more national
stroke clinical registries were identified, information from newly
identified registries were limited. Knowing where important data are
lacking, outdated, or even where a country is ranked might help
facilitate more research or greater policy attention.
Acknowledgments
All
mortality data and some population data were obtained from the World
Health Organization (WHO) mortality database, and the WHO is responsible
only for the provision of the original information. Most population
data were obtained from the United Nations (UN) database, and the UN is
responsible only for the provision of the original information. The
analyses and interpretations of the data are those of the authors alone.
Declaration of conflicting interests
The
author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
Funding
The
author(s) disclosed receipt of the following financial support for the
research, authorship, and/or publication of this article: The authors
received Research Fellowship support from the National Health and
Medical Research Council: D.A.C. (1154273) and A.G.T. (1042600).
ORCID iDs
Tharshanah Thayabaranathan https://orcid.org/0000-0003-2504-7772
Joosup Kim https://orcid.org/0000-0002-4079-0428
Dominique A Cadilhac https://orcid.org/0000-0001-8162-682X
Amanda G Thrift https://orcid.org/0000-0001-8533-4170
Geoffrey A Donnan https://orcid.org/0000-0001-6324-3403
Virginia J Howard https://orcid.org/0000-0003-4912-9975
Valery Feigin https://orcid.org/0000-0002-6372-1740
Mayowa Owolabi https://orcid.org/0000-0003-1146-3070
Jeyaraj Pandian https://orcid.org/0000-0003-0028-1968
Footnote
Authors’ contributions
T.T.,
J.K., and M.T.O. contributed to the design, undertook literature
search, data collection, data analyses and interpretation, wrote the
first draft of the manuscript, and revised the manuscript. D.A.C.,
A.G.T., and G.A.D. contributed to the design, data interpretation, and
revised the manuscript. G.H. contributed to the design, data analyses
and interpretation, and revised the manuscript. V.J.H., P.M.R., V.F.,
B.N., M.O., and J.P. contributed to the design, interpreted the data,
and revised the manuscript. All authors approved the final version of
the manuscript.
No comments:
Post a Comment