2022 - McIntyre - Global Prevalence CP

Received: 24 January 2022 | Accepted: 17 June 2022
DOI: 10.1111/dmcn.15346
ORIGINAL ARTICLE
Global prevalence of cerebral palsy: A systematic analysis
Sarah McIntyre1 | Shona Goldsmith1 | Annabel Webb1 | Virginie Ehlinger2 |

Sandra Julsen Hollung3 | Karen McConnell4 | Catherine Arnaud5 |
1 6 7
Hayley Smithers-Sheedy | Maryam Oskoui | Gulam Khandaker |
8
Kate Himmelmann | on behalf of the Global CP Prevalence Group*
1
Cerebral Palsy Alliance Research Institute,
Specialty of Child and Adolescent Health, Abstract
Sydney Medical School, Faculty of Medicine
and Health, The University of Sydney, Sydney, Aim: To determine trends and current estimates in regional and global prevalence
Australia of cerebral palsy (CP).
2
Center for Epidemiology and Research
in Population health (CERPOP), Inserm, Method: A systematic analysis of data from participating CP registers/surveillance
University of Toulouse, Toulouse, France systems and population-based prevalence studies (from birth year 1995) was per-
3
Norwegian Quality and Surveillance Registry
for Cerebral Palsy (NorCP), Vestfold Hospital formed. Quality and risk of bias were assessed for both data sources. Analyses were
Trust, Tønsberg, Norway conducted for pre-/perinatal, postnatal, neonatal, and overall CP. For each region,
4
School of Health Sciences, Northern Ireland,
UK trends were statistically classified as increasing, decreasing, heterogeneous, or no
5
Public Health Department, University of change, and most recent prevalence estimates with 95% confidence intervals (CI)
Toulouse, Toulouse, France
6
were calculated. Meta-analyses were conducted to determine current birth preva-
Department of Pediatrics, Faculty of
Medicine and Health Sciences, McGill lence estimates (from birth year 2010).
University, Montreal, Canada
7
Results: Forty-one regions from 27 countries across five continents were repre-
Central Queensland Hospital and Health
Service, Rockhampton, Australia sented. Pre-/perinatal birth prevalence declined significantly across Europe and
8
Department of Pediatrics, Institute of Australia (11 out of 14 regions), with no change in postneonatal CP. From the lim-
Clinical Sciences, Sahlgrenska Academy,
University of Gothenburg, Gothenburg, ited but increasing data available from regions in low-and middle-income countries
Sweden
(LMICs), birth prevalence for pre-/perinatal CP was as high as 3.4 per 1000 (95% CI
Correspondence 3.0–3.9) live births. Following meta-analyses, birth prevalence for pre-/perinatal CP
Sarah McIntyre, Cerebral Palsy Alliance
Research Institute, University of Sydney, in regions from high-income countries (HICs) was 1.5 per 1000 (95% CI 1.4–1.6)
Level 4 Building M02C, 88 Mallett Street,
Camperdown, NSW 2050, Australia. live births, and 1.6 per 1000 (95% CI 1.5–1.7) live births when postneonatal CP was
Email: smcintyre@cerebralpalsy.org.au included.
Interpretation: The birth prevalence estimate of CP in HICs declined to 1.6 per
1000 live births. Data available from LMICs indicated markedly higher birth
prevalence.
This original article is commented on by Paneth and Yeargin-Allsopp on pages 1436–1437 of this issue.
Abbreviations: ACPR, Australian Cerebral Palsy Register; HIC, high-i ncome country; LMIC, low-and middle-i ncome country; SCPE, Surveillance of Cerebral Palsy in
Europe.
*Members of the Global CP Prevalence Group are listed in the Acknowledgements.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,
provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2022 The Authors. Developmental Medicine & Child Neurology published by John Wiley & Sons Ltd on behalf of Mac Keith Press.
1494 |
wileyonlinelibrary.com/journal/dmcn Dev Med Child Neurol. 2022;64:1494–1506.
14698749, 2022, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.15346 by Nat Prov Indonesia, Wiley Online Library on [25/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLOBAL PREVALENCE OF CEREBRAL PALSY: A SYSTEMATIC ANALYSIS | 1495
Cerebral palsy (CP) is an umbrella term for a group of disor-

ders of movement and posture, caused by a non-progressive
What this paper adds
interference in the developing brain. Risk factors for CP span
the periods before and around the time of conception, dur- • Birth prevalence of pre-/perinatal cerebral palsy
ing pregnancy, the perinatal period, and up to 2 years of age. (CP) in high-income countries (HICs) is decreasing.
Known risk factors and conditions that can combine into • Current overall CP birth prevalence for HICs is 1.6
causal pathways to CP include genetic variants, congenital per 1000 live births.
anomalies, preterm birth, kernicterus, intrauterine growth • Trends in low-and middle- income countries
restriction and infection, hypoxic ischaemia and cerebrovas- (LMICs) cannot currently be measured.
cular insults during pregnancy and in infancy, and acciden- • Current birth prevalence in LMICs is markedly
tal and non-accidental brain injury.1 higher than in HICs.
Population-based CP registers and prevalence studies • Active surveillance of CP helps to assess the im-
have monitored the birth prevalence of CP for more than pact of medical advancements and social/eco-
60 years.2 The most recent systematic review and meta- nomic development.
analysis of birth prevalence, which mostly included births • Population-based data on prevalence and trends of
in the 1980s and 1990s, found prevalence was 2.1 per 1000 CP are critical to inform policy.
live births.3 Historically, temporal fluctuations have been re-
ported within high-income country (HIC) regions as some
causal pathways become preventable, such as kernicterus,
and others arose such as increased survival of infants born (children living in a region) of CP by region and combined
very preterm with the advent of neonatal intensive care units. for those with data available from birth year 2010 for a cur-
In recent years, significant and sustained declines in the rent prevalence estimate.
birth prevalence of CP in HIC regions of Europe, Australia,
and Japan have been reported.4–7 While the causes for this
decline are complicated, declines are being attributed to M ET HOD
an array of clinical improvements in public health, mater-
nal, and perinatal care, particularly for infants cared for in We sought to maximize the representation of geographi-
a neonatal intensive care unit at highest risk of CP such as cal regions around the world and use the most contempo-
those born very preterm or at term with hypoxic–ischaemic rary data available. We conducted a systematic analysis of
encephalopathy.8 It is therefore important to continue to population-based data from two sources: (1) CP registers/
monitor how improvements in care affect the current birth surveillance systems and (2) published prevalence studies.
prevalence of CP across the world and draw attention to re-
cent trends in CP. It is also important to determine whether
trends are being seen in all HIC regions, as well as to es- Study population
tablish the current overall birth prevalence of CP in these
regions. The study population included children with CP (numera-
Population-based data are also now emerging from re- tor) born from 1995 in regions of the world with population-
gions of low-and middle-income countries (LMICs) where based data, and the population in which they either were
higher rates of CP are being reported.9–11 The aetiologi- born (total live births) or resided (total children of the same
cal pathways to CP in these countries seem to differ from age living in the same region) (denominator). Regions were
HICs.12 As most births worldwide occur in LMICs, it is im- classified by their country's World Bank income classifica-
perative that an update in the prevalence of CP includes data tion (low, lower middle, upper middle, high). Timing of CP
from these regions where possible. was categorized as follows: (1) pre-or perinatal CP—brain
Prevalence of CP is not static and can be expected to con- injury/maldevelopment during the pre-, peri-, or neonatal
tinuously change as a result of medical advancements, and period up to 28 days of life, or unknown aetiology; (2) post-
social and economic development. This study is the result neonatal CP—a known brain-damaging event unrelated to
of an international collaboration which aimed to provide a factors in the ante-, peri-, or neonatal periods, sustained
snapshot of recent changes, and current birth prevalence and after the neonatal period (28 days of life) but before the age
period prevalence (complementary indicators). Specifically, of 2 years; or (3) overall CP—a ll pre-or perinatal CP and
this systematic analysis of CP register data and published postneonatal CP.
literature aimed to identify the following: (1) trends in birth
prevalence for CP of pre-or perinatal origin, postneonatal
CP, and overall CP (live births) by region and combined for CP registers/surveillance systems
two major networks—t he Surveillance of Cerebral Palsy in
Europe (SCPE) and the Australian Cerebral Palsy Register In 2020, invitations to participate were sent to
(ACPR) since birth year 1995; (2) most recent birth preva- representatives from 30 population-based CP registers
lence estimate (live births) and period prevalence estimate known to the study investigators. Registers provided the
1496 | MCINTYRE et al.
aggregated number of children with confirmed CP born/ Sets of two investigators independently reviewed each
living in their region, for each birth year from 1995, by article meeting all eligibility criteria (SG, SM, HSS, SJH,
timing of CP, along with equivalent live birth or population GH, KH, KM). Methods and results data were extracted
denominator data. Data collection was performed during using data extraction sheets designed a priori for the study
2020 to 2021. (including reference, year of publication, geographical loca-
CP was confirmed at a minimum age of 4 years.13 For tion of the study, birth cohorts included, study method, data
the study, the definition of CP14 included the five criteria sources, definition of CP and diagnostic criteria used, age
agreed on by SCPE and the ACPR: (1) is an umbrella term at diagnosis/confirmation of diagnosis, definition and inclu-
for a group of disorders; (2) is a condition that is permanent sion of postneonatal CP, and numerator and denominator
but not unchanging; (3) involves a disorder of movement definitions). The corresponding author was contacted, as
and/or posture and of motor function; (4) is due to a non- required, to clarify information or data, and authors were
progressive interference, lesion, or abnormality, and (5) the asked to provide data from 1995 only. We preferentially
interference, lesion, or abnormality originates in the im- extracted case and denominator data for live births, rather
mature brain.15,16 Registers/surveillance systems providing than children living in the region. When multiple prevalence
data included children with a diagnosis of CP at the age of rates were reported for children at different ages, data were
2 years, but who died before age 4 or 5 years, but excluded extracted for the age group closest to age 5 years, when a di-
children with a diagnosis of CP who died before the age of agnosis of CP is usually confirmed/verified. For aetiological
2 years. timing, CP was categorized as pre-/perinatal if postneonatal
We requested that registers provide descriptive data CP was explicitly excluded; otherwise, data were categorized
about the geographical region represented, size of region, as ‘overall’ CP. If not reported, a denominator was estimated
continuity of data collection, numerator and denominator from the number of cases and prevalence of CP reported and
definitions, definition of CP, data sources and methods of noted in the accompanying tables.
data acquisition, and consent requirements to confirm in-
clusion. To be included in the trends analysis, CP registers/
surveillance systems required a minimum of 10 consecutive Quality and risk of bias assessment
years of data and ongoing data collection.
CP registers/surveillance systems and included publications
were critically appraised for quality and risk of bias. The JBI
Published literature on prevalence of CP checklist for prevalence studies17 was used, which includes
nine quality items (marked as yes, no, unclear, or not appli-
A broad systematic literature search strategy was designed cable) and an overall appraisal to ‘include’ or ‘exclude’ the
with an academic librarian, on the basis of the search origi- study for meta-analysis. Sets of two reviewers independently
nally used by Oskoui et al.3 Searches were conducted in assessed each data source; discrepancies were resolved with an
MEDLINE and EMBASE in November 2020, along with independent third reviewer (SG, SM, HSS, SJH, GH, KH, KM).
handsearching. There were no limits on language of publi-
cation, but the search was limited to papers published from
2011, to include papers published since the systematic review Statistical analysis
by Oskoui et al.3 Abstracts and titles were exported into ref-
erencing software, and automatic and manual de-duplication Objective 1a: recent temporal trends in
was performed (using Covidence systematic review software, each region
Veritas Health Innovation, Melbourne, Australia; available
at www.covidence.org). Titles and abstracts were screened The temporal trend in the number of pre-/perinatal CP cases
for possible inclusion by one investigator (SG), then the full per 1000 live births, and the number of postneonatal CP cases
text of potential articles was retrieved and reviewed by two per 10 000 live births between 1995 and 2014 in each region
(SG and SM). was classified as increasing, decreasing, heterogenous, or no
Original research articles were included if they reported change. For each region, this classification was determined
population-based prevalence of CP from birth year 1995 through a two-step process. First, a Mann–Kendall test15,16 for
with an internationally agreed definition of CP (denomina- monotonic trends was used to determine whether the birth
tors defined as live births or children aged between 0 and prevalence rate for a given region was monotonically increas-
18 years in the region). The following studies were excluded: ing or decreasing. The trend for a given region was classified
(1) abstract available only; (2) studies describing a subgroup as increasing if the resultant Kendall's τ coefficient was posi-
of CP only (e.g. severe motor involvement); (3) studies in- tive and significant (p < 0.05), and was classified as decreas-
cluding people with CP outside the target age range (e.g. ing if the coefficient was negative and significant (p < 0.05).
born before 1995; 50% of children younger than 4 years); and Second, if the trend in a given region could not be classified as
(4) studies from a region already represented in the current either increasing or decreasing (Kendall's τ coefficient giving
study with newer/equivalent data from a participating CP p > 0.05), then a Poisson regression model with an offset term
register/surveillance system or literature. for live births and a smoothing spline term for birth year was
used to distinguish between a heterogenous temporal trend was assessed using the coefficients τ2 and I2 . χ2 tests of het-
and the presence of no change in the birth prevalence rate. erogeneity were also performed. These tests were only used
Cubic B-splines were used for all models, and the degree of descriptively.
smoothness was determined using the restricted maximum Statistical analyses were performed using R version 4.1
likelihood method. The presence of overdispersion in the (packages Kendall version 2.2,18 gam version 1.2,19 ggplot2
Poisson regression was inspected. Regions where the smooth- version 3.3.5,20 meta version 5.2,21 dmetar version 0.0.09,22
ing spline term for birth year was significantly different from mgcv version 1.8;23 R Foundation for Statistical Computing,
zero were classified as heterogenous, while regions where the Vienna, Austria) and STATA version 14.2 (packages
spline term was not significantly different from zero were clas- metaprop_one; StatCorp, College Station, TX, USA).
sified as no change. A smoothed trend line for each region was
plotted to visualize birth prevalence trends.
Ethical review
Objective 1b: combined recent temporal trends In accordance with the National Health and Medical
(register networks) Research Council National Statement on Ethical Conduct
in Human Research (Australia), ethical review was not re-
Temporal trends in the birth prevalence of pre-/perinatal quired for this study as it posed negligible risk and involved
and postneonatal CP using data from two large CP register the use of existing collections of non-identifiable data (tabu-
networks, the SCPE and the ACPR, were analysed by Poisson lated register data and published literature).
regression models with an offset term for live births. Data
from the two networks were pooled after testing for any dif-
ference in trends between them. Orthogonal polynomial R E SU LT S
terms for birth year up to the fourth degree were considered,
and the final form of birth year in the model was selected Data from 41 regions of 27 countries were included. CP reg-
using Akaike information criteria. A quadratic model (poly- isters contributed data representing 19 regions from 15 coun-
nomial up to the second degree) was ultimately selected. tries, all from Europe and Australia, and classified as regions
from HICs. Data from two register networks of Australia and
Europe were also received (ACPR and SCPE). Published lit-
Objective 2a: most recent prevalence in erature provided data from an additional 22 regions from 12
each region countries: Africa (n = 2 regions from LMICs), Asia (n = 4 re-
gions from LMICs and n = 5 regions from HICs), Europe (n = 1
Recent prevalence of pre-/perinatal, postneonatal, and over- region from LMICs and n = 3 from HICs), North America
all CP were calculated for each region, with 95% confidence (n = 7 regions from HICs) (Table 1 and Figure S1). No registers
intervals (CI). Data for the two most recent birth years were or studies were assessed as having a high risk of bias; there-
used for pre-/perinatal CP, while data for any number of fore, all were included in at least one analysis (Tables 1, S1, and
birth years from 2010 were used for postneonatal CP given S2). Data sources not from registers ranged from face-to-face
the small number of individuals with postneonatal CP. The clinical assessments to administrative data linkages (Table 1).
95% CIs were calculated using approximation to the nor- All regions that were able to provide data for trend anal-
mal distribution after proportions were transformed using yses were from HICs. Data for trend analyses were pro-
Freeman–Tukey double arcsine transformation.17 Results are vided by regions from CP registers (13 of 14 regions). The
presented as prevalence rates after the pooled estimates were remaining region from the USA used 1-year survivors as its
back transformed. denominator, and reported results from a surveillance sys-
tem.24 From the 14 regions covering over 8 million live births
that contributed to the pre-/perinatal trend analysis, 79%
Objective 2b: current combined global prevalence showed a statistically significant decline. The regions re-
porting through to 2014 all showed a decline. The remaining
Meta-a nalyses of ‘current’ prevalence of pre- /perinatal three regions showed no change in the time period reported
CP per 1000 live births, postneonatal CP per 10 000 live (Figure 1 and Table S3). However, the most recent data avail-
births, and overall CP per 1000 live births were performed able for the USA were from the early 2000s, the Swiss region
for regions with prevalence data for at least two consecu- represents a very small population, and there have been re-
tive birth years from 2010. The current birth prevalences cent declines in Northern Ireland, but not for the entire pe-
of CP (pre-/perinatal, postneonatal, overall) were derived riod for which data were available for this study (1995–2011).
using univariate meta-analysis of proportions methods. From the 12 regions that were able to provide data for
Random effects meta-a nalyses using the DerSimonian and postneonatal CP, the pattern was mixed, with one region in-
Laird method were preferred to fixed effects meta-a nalysis creasing, one decreasing, three being heterogenous, and seven
because of the anticipated heterogeneity of CP prevalence showing no change. This mixed pattern was also seen in those
between regions. Heterogeneity between regional estimates that provided data through to 2014 (Figure 1 and Table S3).
FIGU R E 1 Birth prevalence trends of cerebral palsy.
There was no difference in the trends between SCPE the area) (Figure 5). Estimated birth years for these studies
and ACPR (test for interaction between birth year and ranged from 1995 to 2016 and covered over 7 million living
register p = 0.67); therefore data were combined for the children (Table S2). Four of the regions were from LMICs and
two register networks. Figure 2 shows the combined trend prevalence ranged from 2.3 to 3.7 per 1000 children. Regions
line and 95% CIs for pre-/perinatal CP across Europe and from HICs ranged from 1.6 to 2.9 per 1000 children; those
Australia with a statistically significant declining trend with higher estimates included much earlier birth years.
(p = 0.012). There was no change for postneonatal CP (fig-
ure not shown).
Birth prevalence estimates with 95% CIs were calculated DISC US SION
for two CP register networks and 25 regions for pre-/perina-
tal CP, 21 regions for postneonatal CP (Figure 3), and 23 re- Before this paper, the most recent international CP birth
gions for overall CP (Figure S2). Most recent birth years were prevalence study was published in 2013.3 As such, the great
included in the analysis; however, they ranged from 1995 to majority of data included were from HICs and birth years
2014. Variation across regions reflect different birth years, from the 1980s and 1990s. At that time the overall birth
size of the denominator population, and World Bank income prevalence was stable, and the estimate was 2.1 per 1000 live
levels. Two regions from LMIC had high birth prevalence es- births. Since then, several studies have been published that
timates of 3 and 3.4 per 1000 live births, one of these regions suggest declines in birth prevalence have occurred in the
also had a high postneonatal CP estimate. 2000s.4,6 This study was therefore undertaken to update our
Meta-analysis was restricted to regions with data since understanding of the global prevalence of CP by using con-
2010 to obtain an estimate for current birth prevalence. A temporary data from CP registers and surveillance systems
total of 17 regions were included in the analysis for both as well as published literature.
pre-/perinatal CP, postneonatal CP, and overall CP, all of Our study confirmed that pre-/perinatal CP is declin-
which were HICs and from CP registers. Heterogeneity does ing in high-income regions in Europe and Australia. The
exist between regions; however, for pre-/perinatal CP there trend was similar for individual regions and for the two
was a combined estimate of 1.5 per 1000 (95% CI 1.4–1.6) major CP register networks, SCPE and ACPR. Only one
live births (τ2 < 0.001, I2 = 69.4%) (Figure 4a). For postneo- other high-i ncome region outside these networks was able
natal CP, the estimate of current birth prevalence was 0.8 to be included;24 however, reporting for this study con-
per 10 000 (95% CI 0.6–1.0) live births (τ2 < 0.001, I2 = 70.1%) cluded in birth year 2002, and the same declining trend
(Figure 4b). For overall CP, the estimate of current birth was not noted. Unfortunately, no registers in LMICs are
prevalence was 1.6 per 1000 (95% CI 1.5–1.7) live births, seen yet able to report on trends, as at least a decade of popula-
in Figure S3 (τ2 < 0.001, I2 = 72.9%). tion data is required to be meaningful. However, a recent
Twelve regions reported period prevalence of CP in chil- systematic review, which used novel methods to predict
dren (denominator being children the same age living in trends, reported a concerning increasing trend in China.25
TA BL E 1 Regions included in the study
Participating in
World Bank temporal trends Participating in most recent birth
Region, country Data or reference Data source income of country Denominator (birth years) prevalence (birth years)
Cross River State, Nigeria Duke et al.36 Key informant method + clinical assessment Lower middle Children — 2003–2014
Eastern Uganda, Uganda Kakooza-Mwesige Door-to-door screening + clinical assessment Low Children — 1998–2013
et al.35
Shahjadpur, Bangladesh Khandaker et al.10 Bangladesh Cerebral Palsy Register Lower middle Live births — 1998–2010
43
Rajshahi Division, Murthy et al. Key informant method + clinical assessment Lower middle Children — 1995–2013
Bangladesh
Henan, China Yuan et al.44 Clinical screening + assessment Upper middle Children — 2005–2010
R.S. Pura Town, India Raina et al.45 Door-to-door screening + clinical assessment Lower middle Children — 1999–2003
46
Japan Toyokawa et al. National health insurance claims High Children — 2004–2009
Okinawa, Japan Touyama et al.7 Okinawa Child Development Center High Live births — 1998–2007
surveillance database
Tochigi, Japan Yamagishi et al.47 Survey from medical record High Live births — 2009–2013
South Korea Park et al.48 National health insurance data High Children — 1999–2003
49
Taiwan Chang et al. National health insurance data High Live births — 1996–2000
GLOBAL PREVALENCE OF CEREBRAL PALSY: A SYSTEMATIC ANALYSIS
Australia Data Australian Cerebral Palsy Register (ACPR) High Live births 1995–2014 2013–2014
New South Wales/ Data NSW/ACT Cerebral Palsy Register High Live births — 2011–2012a
Australian Capital
Territory, Australia
Queensland, Australia Data Queensland Cerebral Palsy Register High Live births — 2010–2011a
South Australia, Data South Australian Cerebral Palsy Register High Live births 1995–2014 2013–2014a
Australia
Victoria, Australia Data Victorian Cerebral Palsy Register High Live births 1995–2014 2013–2014a
Western Australia, Data WA Register of Developmental Anomalies High Live births 1995–2014 2013–2014a
Australia –Cerebral Palsy
Europe Data Surveillance of Cerebral Palsy in Europe High Live births 1995–2010 2009–2010
(SCPE)
Belgium Data Belgian Cerebral Palsy Registry High Live births — 2010–2011a
Croatia Data Register of Cerebral Palsy of Croatia High Live births — 2010–2011a
(RCP-HR)
Denmark Data Danish Cerebral Palsy Registry and National High Live births 1995–2013 2012–2013a
Cerebral Palsy Follow-Up Programme
(CPOP)
Toulouse, France Data Childhood Disabilities Register of the Haute- High Children — 2010–2011
Garonne County (RHE31)
|
1499
(Continues)
TA BL E 1 (Continued)
1500
Participating in
World Bank temporal trends Participating in most recent birth
|
Region, country Data or reference Data source income of country Denominator (birth years) prevalence (birth years)
Grenoble, France Data Register for childhood disabilities and High Children — 2009–2010
perinatal survey (RHEOP)
Attica, Greece Data The Cerebral Palsy Register of Attica High Live births 1999–2011 2010–2011a
Borsod, Hungary Fejes et al.50 Hospital, education, services record review High Live births — 1995–2006
Eastern area, Ireland Data Eastern Area Cerebral Palsy Register High Live births 1995–2012 2011–2012a
Moldova Gincota Bufteac Hospital record review Lower middle Live births — 2009–2010
et al.51
Norway Data Norwegian Quality and Surveillance Registry High Live births 1996–2014 2013–2014a
for Cerebral Palsy (NorCP)
Portugal Data Portuguese Surveillance of Cerebral Palsy High Live births 2001–2011 2010–2011a
Programme
Slovenia Data Slovenian Register of Cerebral Palsy High Live births 1999–2011 2010–2011a
Western Sweden, Sweden Data Cerebral Palsy Register of Western Sweden High Live births 1995–2014 2013–2014a
Saint Gallen, Switzerland Data The Cerebral Palsy Register of St. Gallen High Live births 1995–2011 2010–2011a
(SPRN)
Northern Ireland, UK Data Northern Ireland Cerebral Palsy Register High Live births 1995–2011 2010–2011a
North of England, UK Glinianaia et al.52 North of England Collaborative Cerebral High Live births — 1996–2000
Palsy Survey (NECCPS)
Scotland, UK Bugler et al.53 Cerebral Palsy Integrated Pathway Scotland High Children — 1997–2016
(CPIPS) surveillance program
Sunderland, UK Data Sunderland and area study High Live births 1995–2014 2013–2014a
54
Northern Alberta, Robertson et al. Canadian Cerebral Palsy Registry High Live births — 2008–2010
Canada
Ontario, Canada Ray et al.55 Pre-existing administrative linked dataset High Neonatal — 2002–2008
survivors
Quebec, Canada Oskoui et al.56 Cerebral Palsy Register (REPACQ) High Children — 1999–2001
USA Zablotsky and National health interview survey High Children — 1997–2015
Black 57
Alabama, Georgia, Durkin et al.58 Cerebral Palsy surveillance program (ADDM High Children — 2002
Missouri, Wisconsin, Network)
USA
Metropolitan Atlanta, Van Naarden Cerebral Palsy surveillance program High 1-year survivors 1996–2002 2002
Georgia, USA Braun et al.24 (MADDSP)
South Carolina, USA Li et al.59 Medicaid services, hospital discharge High Live births — 2009
abstracts, department of disabilities/
special needs
a
MCINTYRE et al.
Included in meta-a nalysis of current birth prevalence.
F I G U R E 2 Birth prevalence trend of pre-/perinatal cerebral palsy (CP; Surveillance of Cerebral Palsy in Europe and the Australian Cerebral Palsy
Register combined).
Neonatal intensive care units are expanding in China and advances in health care may lead to increases in CP preva-
other LMICs, and the increased survival of medically frag- lence, as well as decreases.
ile infants born at increasingly lower gestational ages may The number of CP registers and prevalence studies in
result in an initial spike in CP prevalence, while further LMICs is increasing, yet they remain extremely under-
development of neonatal care may decrease the prevalence represented. In this study, 7 out of 41 regions were from
again—as has been seen in HICs.26 A challenge for all is LMICs (Nigeria, Uganda, Bangladesh [n = 2], China, India,
how to share knowledge, experiences, and lessons learnt to Moldova) compared with 3 out of 49 in the previous study
minimize this inevitable spike. A recent large randomized (China, Kenya, Turkey).3 Two regions were able to report
controlled trial of therapeutic hypothermia has shown that birth prevalence using live births as a denominator, making
we cannot assume that standard interventions in HICs will this comparable to the high-income regions. Birth preva-
work in the same way in LMIC settings;27 evaluation of lence was 3.3 per 1000 overall for Shahjadpur, Bangladesh,
such interventions is essential before being introduced into and 3.4 per 1000 for pre-/perinatal CP in Moldova. These
new settings and prevention opportunities should remain birth prevalence estimates are more than double the find-
a priority.28 ings for high-income regions in our meta-a nalysis. The re-
To calculate current global birth prevalence estimates, we mainder reported period prevalence (with a denominator of
restricted meta-analyses to regions with more than one birth children living in the region) as high as 3.7 per 1000 chil-
year from 2010. No LMICs were able to participate in these dren in Rajshahi Division, Bangladesh. Additional litera-
analyses, so these primary findings are for high-income re- ture from Albania, Egypt, and Pakistan, which could not be
gions only. The current pre-/perinatal CP birth prevalence is included in this review, supports these findings.32–34 Low-
1.5 per 1000 live births. The current overall (including post- and middle-income regions reporting prevalence estimates
neonatal) CP birth prevalence is 1.6 per 1000 live births. This suggest that these are almost certainly underestimates due
prevalence estimate is 25% lower than the overall birth prev- to survival bias (i.e. high mortality in the early years, before
alence estimate reported in 2013 (2.1 per 1000),3 and this a CP diagnosis and, again, high mortality in children with
updated current birth prevalence estimate for HICs should CP), incomplete ascertainment, and inability to include
now be used. This is particularly encouraging as this decline very mild cases at population level.10,35,36 Collaborative ef-
has occurred during the same era that survival in neonatal forts such as mentorship programmes with SCPE, ACPR,
intensive care units is improving for infants born extremely and the Global LMIC CP register will increase representa-
preterm.31 As described earlier, we have learnt to expect that tion of LMICs.37
F I G U R E 3 Regional birth prevalence of cerebral palsy. ACPR, Australian Cerebral Palsy Register; ACT, Australian Capital Territory; NSW, New
South Wales; SCPE, Surveillance of Cerebral Palsy in Europe.
There has been no change in postneonatal CP, and the differences in aetiologies of postneonatal CP (e.g. malaria,
current estimate for HICs is 0.8 per 10 000 live births with previous nutritional status) and potentially different oppor-
wide confidence intervals. The numbers for postneonatal tunities for prevention strategies that are specific to each
CP in HICs are small, and we have less confidence about region.37,38
these trends, particularly for children with a brain injury A shared understanding of the definition and classifi-
closer to the age of 2 years, which may be described as an cation of CP is essential for reliable estimates and trends.
acquired brain injury rather than postneonatal CP. LMICs Standardized and consistent approaches used by registers
with higher proportions of postneonatal CP (up to 36% in enable accurate monitoring of the condition over time. In
Nigeria,36 compared with 6% in Australia2), alert us to the situations when complete agreement cannot be reached, data
F I G U R E 4 (a) Current pre-/perinatal birth prevalence of cerebral palsy (CP) in high-income countries. (b) Current post neonatal birth prevalence of
CP in high-income countries. ACT, Australian Capital Territory; CI, confidence interval; NSW, New South Wales.
can be harmonized for comparisons.39 For example, in this prevalence is a relevant indicator of the impact of the organi-
study, data were restricted to CP that occurred in the first zation of care and practices in the peri-and neonatal period,
2 years of life, and children who survived beyond the age of the cross-sectional approach used for period prevalence
2 years, despite variations between CP registers in these lim- estimates is more relevant for documenting public health
its.13 We recommend the continued use of papers such as the issues, notably the impact of CP in the community. It is gen-
one by Smithers-Sheedy et al.13 (including confirming diag- erally accepted that period prevalence is higher than birth
nosis at age 4 years) and the full annotation that describes in prevalence, which is consistent with our results, although
detail the definition of CP.14 comparisons are difficult (small sample size and different
A strength of this study was our reporting on both birth regions). Another strength of this study was our representa-
prevalence and period prevalence, which is rarely done. tion of regions without CP registers, by including published
These two indicators are complementary. While birth population-based studies with alternative methodologies,
FIGU R E 5 Period prevalence of overall cerebral palsy for children.
such as large surveys and administrative data. However, it Gergeli, Catherine Gibson, Karen Horridge, Christoph
is known that conditions, including CP, are coded incon- Tobias Kuenzle, Svetlana V Glinianaia, and Malika Delobel-
sistently,40 so risk of bias was higher for studies that solely Ayoub. Open access publishing facilitated by The University
relied on administrative data for identifying children with of Sydney, as part of the Wiley -The University of Sydney
CP. Finally, we observed regional heterogeneity in our birth agreement via the Council of Australian University
prevalence meta-analyses. Although there will always be a Librarians. Open access publishing facilitated by The
level of true variation in prevalence between regions, it is University of Sydney, as part of the Wiley - The University of
likely that under-ascertainment of children with mild CP is Sydney agreement via the Council of Australian University
also a contributing factor, particularly for newly established Librarians.
registers.41 We thank the regions that contributed data to the study.
The declining trends in the birth prevalence of pre-/ We sincerely thank all the families whose data make the
perinatal CP, evidenced by CP registers in this paper, in- epidemiological research possible. The authors have stated
creases our understanding of the condition and the im- that they had no interests that might be perceived as posing
pact of improvements in ante-, peri-, and postnatal care in a conflict or bias.
HICs. This global overview represents the recent and cur-
rent situation in over 40 regions of the world. Sustainable F U N DI N G I N F OR M AT ION
registers with good ascertainment are essential for con- The Australian CP Register and the NSW/ACT CP Register
tinued monitoring of trends and prevalence, and the are funded by the Cerebral Palsy Alliance, Australia. The
real-world impact of changing social development and Northern Ireland Cerebral Palsy Register is funded by the
health care in low-, middle-, and high-i ncome countries. Public Health Agency Northern Ireland. The Norwegian
Recognition of CP at national and international levels Quality and Surveillance Registry for Cerebral Palsy is
provides a powerful tool to potentially influence policy funded by the South- Eastern Norway Regional Health
and services, leading to a demonstrable contribution to Authority. The Registry of Childhood Disabilities in Haute-
society and economies.42 Garonne county (RHE31), France, is funded by the Public
Health Agency France and the National Institute of Health
AC K N OW L E D G M E N T S and Medical Research. Registre des Handicaps de l'Enfant et
The additional members of the Global CP Prevalence Group Observatoire Périnatal, Grenoble, France.
are as follows: Gina Hinwood, Linda Watson, Megan Auld,
Natasha Garrity, Nadia Badawi, Els Ortibus, Inge Franki, DATA AVA I L A BI L I T Y S TAT E M E N T
Vlatka Mejaski-Bosnjak, Gija Rackauskaite, Elodie Sellier, Data sharing is available on request to the authors, and if
Antigone Papavasileiou, Melinda Fejes, Valerie Dowding, data is from a register it would require approval from the
Claire Kerr, Guro L. Andersen, Daniel Virella, Anja Troha individual register.
18. McLeod AI. Kendall: Kendall rank correlation and Mann-Kendall

ORC I D trend test. R package version 2.2. https://CRAN.R-project.org/pack-
Sarah McIntyre https://orcid.org/0000-0002-0234-1541 age = Kendall. 2011.
Shona Goldsmith https://orcid.org/0000-0003-3903-6142 19. Hastie T. gam: Generalized Additive Models. R package version 1.20.
Annabel Webb https://orcid.org/0000-0001-8435-4436 https://cran.r-project.org/web/packages/gam/gam.pdf. 2020.
20. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-
Virginie Ehlinger https://orcid.org/0000-0003-4992-5998
Verlag New York; 2016.
Sandra Julsen Hollung https://orcid.org/0000-0002-7486-7454 21. Balduzzi S, Rücker G, Schwarzer G. How to perform a meta-analysis with
Karen McConnell https://orcid.org/0000-0002-5221-9800 R: a practical tutorial. Evidence-Based Mental Health 2019; 22: 153–60.
Catherine Arnaud https://orcid.org/0000-0002-4002-802X 22. Harrer M, Cuijpers P, Furukawa T, et al. dmetar: Companion R
Hayley Smithers-Sheedy https://orcid. Package For The Guide ‘Doing Meta-A nalysis in R'. R package version
0.0.9000. http://dmetar.protectlab.org/. 2019.
org/0000-0002-0082-2413
23. Wood SN. Generalized Additive Models: An Introduction with R
Maryam Oskoui https://orcid.org/0000-0003-1042-0108 (2nd edition). New York: Chapman and Hall/CRC; 2017.
Gulam Khandaker https://orcid. 24. Van Naarden Braun K, Doernberg N, Schieve L, et al. Birth Prevalence
org/0000-0002-0661-4113 of Cerebral Palsy: A Population-Based Study. Pediatrics 2016; 137: 1–9.
Kate Himmelmann https://orcid. 25. Yang S, Xia J, Gao J, et al. Increasing prevalence of cerebral palsy
among children and adolescents in China 1988-2020: A systematic
org/0000-0002-3959-9554
review and meta-a nalysis. J Rehabil Med 2021; 53: jrm00195.
26. Faruk T, King C, Muhit M, et al. Screening tools for early identifica-
R EFER ENCES tion of children with developmental delay in low-and middle-income
1. McIntyre S, Morgan C, Walker K, et al. Cerebral palsy--don't delay. countries: a systematic review. BMJ Open 2020; 10: e038182.
Dev Disabil Res Rev 2011; 17: 114–29. 27. Thayyil S, Pant S, Montaldo P, et al. Hypothermia for moderate or
2. Himmelmann K, McIntyre S, Goldsmith S, et al. Epidemiology of severe neonatal encephalopathy in low-income and middle-income
Cerebral Palsy. In: Miller F, Bachrach S, Lennon N, O'Neil M editors. countries (HELIX): a randomised controlled trial in India, Sri Lanka,
Cerebral Palsy. Switzerland: Springer, Cham; 2018. and Bangladesh. Lancet Glob Health 2021; 9: e1273-e85.
3. Oskoui M, Coutinho F, Dykeman J, et al. An update on the preva- 28. Thayyil S, Bassett P, Shankaran S. Questions about the HELIX trial
lence of cerebral palsy: a systematic review and meta-a nalysis. Dev -Authors' reply. Lancet Glob Health 2021; 9: e1654-e5.
Med Child Neurol 2013; 55: 509–19. 29. Hollung SJ, Vik T, Lydersen S, et al. Decreasing prevalence and se-
4. Galea C, McIntyre S, Smithers-Sheedy H, et al. Cerebral palsy trends verity of cerebral palsy in Norway among children born 1999 to 2010
in Australia (1995-2009): a population-based observational study. concomitant with improvements in perinatal health. Eur J Paediatr
Dev Med Child Neurol 2019; 61: 186–93. Neurol 2018; 22: 814–21.
5. Reid SM, Meehan E, McIntyre S, et al. Temporal trends in cerebral 30. Larsen ML, Rackauskaite G, Greisen G, et al. Continuing decline in
palsy by impairment severity and birth gestation. Dev Med Child the prevalence of cerebral palsy in Denmark for birth years 2008-
Neurol 2016; 58 25–35. 2013. Eur J Paediatr Neurol 2020; 30: 155–61.
6. Sellier E, Platt MJ, Andersen GL, et al. Decreasing prevalence in ce- 31. Stoll BJ, Hansen NI, Bell EF, et al. Trends in Care Practices, Morbidity,
rebral palsy: a multi-site European population-based study, 1980 to and Mortality of Extremely Preterm Neonates, 1993-2012. JAMA
2003. Dev Med Child Neurol 2016; 58: 85–92. 2015; 314: 1039–51.
7. Touyama M, Touyama J, Toyokawa S, et al. Trends in the prevalence 32. Kruja J, Beghi E, Zerbi D, et al. High prevalence of major neurologi-
of cerebral palsy in children born between 1988 and 2007 in Okinawa, cal disorders in two Albanian communities: results of a door-to-door
Japan. Brain Dev 2016; 38: 792–9. survey. Neuroepidemiology 2012; 38: 138–47.
8. Badawi N, McIntyre S, Hunt RW. Perinatal care with a view to pre- 33. El-Tallawy HN, Farghaly WM, Shehata GA, et al. Epidemiology of
venting cerebral palsy. Dev Med Child Neurol 2021; 63: 156–61. cerebral palsy in El-K harga District-New Valley (Egypt). Brain Dev
9. Donald KA, Samia P, Kakooza-Mwesige A, et al. Pediatric cerebral 2011; 33: 406–11.
palsy in Africa: a systematic review. Seminars in pediatric neurology 34. El-Tallawy HN, Farghaly WM, Shehata GA, et al. Cerebral palsy
2014; 21: 30–5. in Al-Quseir City, Egypt: prevalence, subtypes, and risk factors.
10. Khandaker G, Muhit M, Karim T, et al. Epidemiology of cerebral Neuropsychiatr Dis Treat 2014; 10: 1267–72.
palsy in Bangladesh: a population-based surveillance study. Dev Med 35. Kakooza-Mwesige A, Andrews C, Peterson S, et al. Prevalence of ce-
Child Neurol 2019; 61: 601–9. rebral palsy in Uganda: a population-based study. The Lancet Global
11. Kakooza-Mwesige A, Andrews C, Peterson S, et al. Prevalence of ce- Health 2017; 5: e1275-e82.
rebral palsy in Uganda: a population-based study. Lancet Glob Health 36. Duke R, Torty C, Nwachukwu K, et al. Clinical features and aetiology
2017; 5: e1275-e82. of cerebral palsy in children from Cross River State, Nigeria. Archives
12. Khandaker G, Smithers-Sheedy H, Islam J, et al. Bangladesh Cerebral of Disease in Childhood 2020; 105: 625–30.
Palsy Register (BCPR): a pilot study to develop a national cerebral 37. Jahan I, Muhit M, Hardianto D, et al. Epidemiology of cerebral palsy
palsy (CP) register with surveillance of children for CP. BMC Neurol in za\ies: preliminary findings from an international multi-centre ce-
2015; 15: 173. rebral palsy register. Dev Med Child Neurol 2021; 63: 1327–36.
13. Smithers-Sheedy H, Badawi N, Blair E, et al. What constitutes cere- 38. Jahan I, Al Imam MH, Karim T, et al. Epidemiology of cerebral
bral palsy in the twenty-first century? Dev Med Child Neurol 2014; palsy in Sumba Island, Indonesia. Dev Med Child Neurol 2020; 62:
56: 323–8. 1414–22.
14. Rosenbaum P, Paneth N, Leviton A, et al. A report: the definition and 39. Sellier E, McIntyre S, Smithers- Sheedy H, et al. European and
classification of cerebral palsy April 2006. Dev Med Child Neurol Australian Cerebral Palsy Surveillance Networks Working Together
Suppl 2007; 109: 8–14. for Collaborative Research. Neuropediatrics 2020; 51: 105–12.
15. Mann HB. Nonparametric tests against trend. Econometrica: Journal 40. The National Confidential Enquiry into Patient Outcome and Death.
of the econometric society 1945; 13: 245–59. Each and Every Need. London: National Confidential Enquiry into
16. Kendall MG. Rank Correlation Methods, 4th edition. London: Patient Outcome and Death; 2018.
Charles Griffin; 1975. 41. Goldsmith S, McIntyre S, Smithers-Sheedy H, et al. An international
17. Freeman MF, Tukey JW. Transformations related to the angular and survey of cerebral palsy registers and surveillance systems. Dev Med
the square root. The Annals of Mathematical Statistics 1950: 607–11. Child Neurol 2016; 58 Suppl 2: 11–7.
42. UK Research and Innovation: Economic and Social Research Council. 56. Oskoui M, Joseph L, Dagenais L, et al. Prevalence of cerebral palsy
Defining Impact. https://www.ukri.org/counci ls/esrc/impact-toolk in Quebec: alternative approaches. Neuroepidemiology 2013; 40:
it-for-economic-a nd-socia l-sciences/defining-impact/ (accessed 264–8 .
03/12/2021 57. Zablotsky B, Black LI. Prevalence of Children Aged 3-17 Years With
43. Murthy GV, Mactaggart I, Mohammad M, et al. Assessing the preva- Developmental Disabilities, by Urbanicity: United States, 2015-2018.
lence of sensory and motor impairments in childhood in Bangladesh National health statistics reports 2020; 139: 1–7.
using key informants. Arch Dis Child 2014; 99: 1103–8. 58. Durkin MS, Benedict RE, Christensen D, et al. Prevalence of Cerebral
44. Yuan J, Wang J, Ma J, et al. Paediatric cerebral palsy prevalence and Palsy among 8-Year-Old Children in 2010 and Preliminary Evidence
high-risk factors in Henan province, Central China. Journal of reha- of Trends in Its Relationship to Low Birthweight. Paediatric and
bilitation medicine 2019; 51: 47–53. Perinatal Epidemiology 2016; 30: 496–510.
45. Raina SK, Razdan S, Nanda R. Prevalence of cerebral palsy in chil- 59. Li Q, Kinsman SL, Jenkins DD, et al. Decreasing prevalence of cere-
dren < 10 years of age in R.S. Pura town of Jammu and Kashmir. bral palsy in birth cohorts in South Carolina using Medicaid, disabil-
Journal of Tropical Pediatrics 2011; 57: 293–5. ity service, and hospital discharge data, 1996 to 2009. Developmental
46. Toyokawa S, Maeda E, Kobayashi Y. Estimation of the number of chil- Medicine & Child Neurology 2019; 61: 593–600.
dren with cerebral palsy using nationwide health insurance claims
data in Japan. Developmental Medicine & Child Neurology 2017; 59:
317–21. S U PP ORT I N G I N F OR M AT ION
47. Yamagishi H, Osaka H, Toyokawa S, et al. Survey on Children with
The following additional material may be found online:
Cerebral Palsy in Tochigi Prefecture, Japan. Pediatrics International
2020; 11: 11. Figure S1: Data sources flow chart.
48. Park MS, Kim SJ, Chung CY, et al. Prevalence and lifetime health- Figure S2: Overall regional birth prevalence of cerebral
care cost of cerebral palsy in South Korea. Health Policy 2011; 100: palsy.
234–8. Figure S3: Current overall birth prevalence of cerebral palsy
49. Chang MJ, Ma HI, Lu TH. Estimating the prevalence of cerebral palsy
in high-income countries.
in Taiwan: A comparison of different case definitions. Research in
Developmental Disabilities 2015; 36: 207–12. Appendix S1: Search strategy.
50. Fejes M, Varga B, Hollody K. [Epidemiology, cost and economic im- Table S1: JBI/supporting data for regions represented in
pact of cerebral palsy in Hungary]. Ideggyogy Sz 2019; 72: 115–22. birth prevalence analyses.
51. Gincota Bufteac E, Andersen GL, Torstein V, et al. Cerebral palsy Table S2: JBI/supporting data for regions represented in
in Moldova: subtypes, severity and associated impairments. BMC
period prevalence analyses.
Pediatrics 2018; 18: 332.
52. Glinianaia SV, Rankin J, Colver A. Cerebral palsy rates by birth Table S3: Supporting data for regional trends in pre- /
weight, gestation and severity in North of England, 1991-2000 single- perinatal cerebral palsy and postneonatal cerebral palsy.
ton births. Archives of Disease in Childhood 2011; 96: 180–5.
53. Bugler KE, Gaston MS, Robb JE. Distribution and motor ability of
children with cerebral palsy in Scotland: a registry analysis. Scottish How to cite this article: McIntyre S, Goldsmith S,
Medical Journal 2019; 64: 16–21.
Webb A, et al. Global prevalence of cerebral palsy: A
54. Robertson CMT, Florencia Ricci M, O'Grady K, et al. Prevalence
Estimate of Cerebral Palsy in Northern Alberta: Births, 2008-2010. systematic analysis. Dev Med Child Neurol.
Canadian Journal of Neurological Sciences 2017; 44: 366–74. 2022;64(12):1494–1506. https://doi.org/10.1111/
55. Ray JG, Redelmeier DA, Urquia ML, et al. Risk of cerebral palsy dmcn.15346
among the offspring of immigrants. PLoS One 2014; 9: e102275.

2022 - McIntyre - Global Prevalence CP

Uploaded by

Copyright:

Available Formats

2022 - McIntyre - Global Prevalence CP

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2022 - McIntyre - Global Prevalence CP

Uploaded by

Copyright:

Available Formats

Received: 24 January 2022 | Accepted: 17 June 2022

Global prevalence of cerebral palsy: A systematic analysis

Sarah McIntyre1 | Shona Goldsmith1 | Annabel Webb1 | Virginie Ehlinger2 |

Cerebral palsy (CP) is an umbrella term for a group of disor-

FIGU R E 1 Birth prevalence trends of cerebral palsy.

Included in meta-a nalysis of current birth prevalence.

FIGU R E 5 Period prevalence of overall cerebral palsy for children.

18. McLeod AI. Kendall: Kendall rank correlation and Mann-Kendall

You might also like

2022 - McIntyre - Global Prevalence CP

Uploaded by

Copyright:

Available Formats

2022 - McIntyre - Global Prevalence CP

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2022 - McIntyre - Global Prevalence CP

Uploaded by

Copyright:

Available Formats

Received: 24 January 2022 | Accepted: 17 June 2022

Global prevalence of cerebral palsy: A systematic analysis

Sarah McIntyre1 | Shona Goldsmith1 | Annabel Webb1 | Virginie Ehlinger2 |

Cerebral palsy (CP) is an umbrella term for a group of disor-

FIGU R E 1 Birth prevalence trends of cerebral palsy.

Included in meta-­a nalysis of current birth prevalence.

FIGU R E 5 Period prevalence of overall cerebral palsy for children.

18. McLeod AI. Kendall: Kendall rank correlation and Mann-­Kendall

You might also like

Included in meta-a nalysis of current birth prevalence.

18. McLeod AI. Kendall: Kendall rank correlation and Mann-Kendall