Global Burden of Childhood Epilepsy, Intellectual Disability, and Sensory Impairments

The GBD provides estimates of non-fatal outcomes for “impairments” (used interchangeably with “disabilities” in the current article), as part of the annual comprehensive assessment of incidence, prevalence, and YLD for several health conditions across 195 countries and territories. Impairments are defined as sequelae of multiple causes for which better data are available to estimate the overall occurrence than for each underlying cause. Four such impairments are developmental intellectual disability, epilepsy, hearing loss, and vision loss. Case definitions and diagnostic criteria for these impairments were based on International Classification of Diseases, Ninth Revision (ICD-9) and International Classification of Diseases, 10th Revision (ICD-10) codes, complemented with relevant guidelines such as the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition¹⁹  and the Guidelines for Epidemiologic Studies on Epilepsy.²⁰  A detailed description of the GBD conceptual framework for epilepsy is reported elsewhere.²¹  Disorders of intellectual development in the ICD-9 and ICD-10 codes are termed as “developmental intellectual disability” or simply “intellectual disability” in this article and were grouped into 5 bands on the basis of IQ scores: borderline, mild, moderate, severe, and profound intellectual disability. Hearing and visual impairments were similarly classified into bands of severity corresponding to frequency response and visual acuity cutoffs, respectively. These 4 disabilities were selected for comparability with the GBD 2004 estimates for childhood disabilities.

The detailed methodologic techniques used for estimating the burden of each impairment have been reported previously^16–18  and are presented in the Supplemental Information. In brief, the estimation for each condition started with the compilation of all available data inputs from systematic reviews of the literature, hospital and claims databases, health surveys, case notification systems, cohort studies, and multinational survey data. All input data for GBD 2017 are available at the Global Health Data Exchange (http://ghdx.healthdata.org/gbd-2017/data-input-sources). A total of 1675 sources were analyzed for data on the prevalence of intellectual disability (57), hearing loss (355), epilepsy (455), and vision loss (808). Efforts were made to (1) optimize the comparability of data derived from various sources by using different methods, (2) find a consistent set of estimates across prevalence data, and (3) generate estimates for locations with sparse or no data by using available information from other locations combined with covariates. Prevalence estimates were produced by using DisMod-MR 2.1, a statistical modeling technique developed for the GBD project (Supplemental Fig 5).²²  This is a Bayesian meta-regression tool that synthesizes epidemiological data for fatal and non-fatal health outcomes from disparate settings and sources, adjusting for different case definitions, diagnostic criteria, or sampling methods, to generate internally consistent estimates by geographical location, year, age group, and sex. The validity of DisMod has been widely reported.^16–18,22 

YLD, defined as the years of life lived with a condition in a less than ideal health state, are designed to provide a comparable measure of disease burden across diverse health conditions and impairments rather than a measure of functional status, as described in ICF.¹⁵  To calculate YLD for the 4 disabilities, the estimated prevalence of each disability was multiplied by an assigned disability weight. Disability weights are the population assessment of magnitude of health loss from specific health outcomes measured on a scale from 0 to 1, in which 0 equals a state of perfect health and 1 equals death. The disability weights were estimated from multicountry population-based surveys, as described in detail elsewhere.^18,23 

The global estimates of the prevalence and YLD for the 4 disabilities in children and adolescents were disaggregated by age group (children: <1, 1–4, 5–9 years; adolescents: 10–14 and 15–19 years), sex, and geographical regions (high-income North America, Western Europe, Central and Eastern Europe and Central Asia, Latin America and the Caribbean, Southeast and East Asia and Oceania, South Asia, Sub-Saharan Africa, and North Africa and the Middle East). The selected locations are not geopolitical units, but groupings of countries created by GBD for analytical purposes. All computations in the GBD were conducted 1000 times to propagate uncertainty around the estimates for prevalence and YLD. At every step in the modeling process, the distributions were assessed for the sampling error of data inputs, the uncertainty of data corrections for measurement errors, the uncertainty in coefficients from model fit, and the uncertainty of severity distributions and disability weights. The corresponding uncertainty intervals (UIs) for prevalence and YLD estimates of the 4 disabilities were defined at the 2.5th and 97.5th value of 1000 draws. As with all GBD articles, the substantive data that formed the basis of this analysis^16,17  adhered to the Guidelines for Accurate and Transparent Health Estimates Reporting, which include recommendations on the documentation of data sources, estimation methods, statistical analysis, and statistical code.²⁴ 

Globally, the population of children and adolescents was ∼2.6 billion in 2017, of whom 291.2 million (95% UI: 249.9–335.4), or 11.2% (95% UI: 10.0–12.5), were estimated to have 1 of the 4 disabilities examined (Tables 1–4). Approximately 152.3 million (52.3%) were male, although the sex pattern varied across the disabilities. The prevalence of these disabilities increased with age, from 6.1% among all the ∼138 million children aged <1 year to 13.9% among the roughly 616 million adolescents aged 15 to 19 years. A total of 16 million (95% UI: 13.4–18.9 [5.5%]) lived in high-income countries, and 275.2 million (95% UI: 236.4–316.5 [94.5%]) lived in LMIC. Of all the children and adolescents with disabilities, 8.4 million (2.9%) were aged <1 year, 47.9 million (16.4%) were aged 1 to 4 years, 73.5 million (25.2%) were aged 5 to 9 years, 75.9 million (26.1%) were aged 10 to 14 years, and 85.6 million (29.4%) were aged 15 to 19 years. Thus, a total of 205.6 million (70.6%) were <15 years of age. These 4 disabilities accounted for 28.9 million YLD (or 19.9%) of the overall 145.3 million (95% UI: 106.9–189.7) YLD among children and adolescents from all causes of fatal and non-fatal outcomes included in the GBD 2017.

The prevalence of hearing loss rose from 0.9% among children aged <1 year to 5.9% (95% UI: 5.4%–6.5%) among adolescents aged 15 to 19 years (Fig 1). Vision loss rose from 1.1% among children aged <1 year to 3.9% (95% UI: 3.4%–4.6%) among adolescents aged 15 to 19 years. The prevalence of intellectual disability and epilepsy remained largely constant at ∼3% and 0.9%, respectively, in all age groups. Among all children and adolescents, the disability-specific prevalence was 0.9% (95% UI: 0.8%–1.1%) for epilepsy, 3.2% (95% UI: 2.5%–3.9%) for intellectual disability, 3.1% (95% UI: 2.7%–3.6%) for vision loss, and 4.0% (95% UI: 3.7%–4.3%) for hearing loss. However, epilepsy and intellectual disability were associated with the highest YLD in all age groups, which were significantly higher than sensory disabilities.

South Asia accounted for the highest prevalence of intellectual disability (6.0% [95% UI: 4.5%–7.5%]), hearing loss (4.7% [95% UI: 4.3%–5.0%]), and vision loss (3.7% [95% UI: 3.2%–4.2%]), whereas Latin America and the Caribbean recorded the highest prevalence of epilepsy (1.2% [95% UI: 1.0%–1.5%]), as shown in Fig 2. Epilepsy was least prevalent in Southeast and East Asia and Oceania (0.7% [95% UI: 0.6%–0.8%]), and intellectual disability was least prevalent in Latin America and the Caribbean (1.6% [95% UI: 1.3%–1.9%]), whereas vision loss (1.2% [95% UI: 1.0%–1.5%]) and hearing loss (2.2% [95% UI: 2.1%–2.4%] were least prevalent in high-income North America. South Asia (0.6% [95% UI: 0.4%–0.7%]) and Latin America and the Caribbean (0.5% [95% UI: 0.4%–0.7%]) were associated with the highest YLD rates for intellectual disability and epilepsy, respectively. South Asia (107.8 million or 37.5%) and sub-Saharan Africa (59.8 million or 20.5%) accounted for more than half of all children and adolescents with disabilities (Fig 3). High-income North America (5.6 million or 1.9%) and Western Europe (6.2 million or 2.1%) accounted for the lowest population of children and adolescents with disabilities.

Globally, mild intellectual disability (IQ scores of 50–69) was the most prevalent disability, and mild vision loss was the least prevalent disability among the 4 types of disability (Fig 3). Within regions, mild hearing loss was the most common disability in all regions except Central and Eastern Europe and Central Asia and Southeast and East Asia and Oceania. Mild vision loss was the least prevalent disability in all regions. Severe epilepsy and severe intellectual disability were the 2 most common disabilities with the highest YLD globally and in all regions. The burden of epilepsy was more concentrated in Mexico, Colombia, and Venezuela in Latin America and was more concentrated in Gabon, the Republic of the Congo, and Angola in sub-Saharan Africa (Fig 4). Intellectual disability had the highest prevalence in India (South Asia), Afghanistan, and Yemen (Middle East). Hearing loss was most prevalent in Madagascar (sub-Saharan Africa) and Myanmar (Southeast Asia), whereas vision loss was more widespread and had the highest prevalence in South Sudan, the Central African Republic (sub-Saharan Africa), and Papua New Guinea (Oceania).

The top 10 countries accounted for 52.8% of all children and adolescents with epilepsy, 68.0% of children and adolescents with intellectual disability, 62.0% of children and adolescents with hearing loss, and 59.4% of children and adolescents with vision loss (Supplemental Fig 6). The top 10 countries accounted for 62.3% of all children and adolescents with these disabilities. These countries also accounted for at least 53.5% of the YLD associated with these conditions. India and China had the highest population of children and adolescents with any disability, with their associated YLD. The United States was the only high-income country among the top 10 countries for any disability. However, in terms of the highest prevalence and YLD rates per population, countries with the highest prevalence were Gabon for epilepsy, India for intellectual disability, Madagascar for hearing loss, and Israel for vision loss. The underlying causes or risk factors and their ICD-9 and ICD-10 codes that formed the basis of the reported estimates for all disabilities in the GBD 2017, as well as their contributions to the aggregate estimate for each disability, are summarized in Supplemental Fig 7.

Our primary aim was to update the global estimate of children and adolescents with disabilities previously attributed to the GBD 2004 and still widely reported in the literature.^2–9  Although, in our study, we did not capture the full spectrum of all possible childhood disabilities, our analysis reveals that the number of children and adolescents with disabilities is at least 291 million globally and that the prevalence increases with age. Among children and adolescents <15 years, the estimate of children and adolescents with these 4 disabilities is more than twice the 2004 estimate of 93 million. The substantially higher estimate can be attributed to several factors besides the modest impact of the ∼6.6% and 4.1% rise in the global population of children and adolescents <15 and <20 years, respectively, between 2004 and 2017.²⁵  Firstly, the number and variety of data sources used in generating estimates have increased substantially since 2004. Secondly, the modeling techniques have improved significantly on the basis of expert contributions from an increasing number of institutional and individual collaborators from 146 countries.²⁶  Thirdly, consistent with the classification recommended by the ICF, the GBD 2017 estimates include disabilities of varying degrees of severity from slight or mild to profound for each disability. Lastly, there is now an internationally agreed on framework for evaluating estimates of health from statistical modeling, which has been strictly followed by the GBD since 2015.²⁴ 

The GBD modeling efforts are meant to bridge a critical gap in the epidemiology of developmental disorders resulting from several conceptual and operational challenges in measuring disabilities in children and adolescents.^27,28  As with most health conditions, the dearth of population-based data for specific disabilities, especially in LMIC, has compelled a growing reliance on the statistical estimation of health outcomes as an interim step to guide health policy and intervention. The implicit philosophy underpinning these efforts is that the absence of conventional epidemiological data is not evidence of the absence of a health condition. Thus, GBD is an essential, independent, and up-to-date source of alternative data for policy-makers and decision-makers in global health, especially for countries with poor or no data.^2,3 

However, it is helpful to evaluate the plausibility of the GBD estimates in the absence of comparable global estimates from other reputable sources. The global estimate of 0.9% (95% UI: 0.8%–1.1%) for epilepsy among all children and adolescents appears plausible on the basis of evidence in the literature, suggesting a global prevalence of between 0.5% and 1%.^29,30  In addition, the estimate of 1.0% (95% UI: 0.8%–1.3%) for sub-Saharan Africa is consistent with the 0.9% reported in the most comprehensive systematic review on epilepsy to date from this region.³¹  The GBD estimate for intellectual disability of 3.2% (95% UI: 2.5%–3.9%) was higher than the estimate of 1.8% reported in 1 meta-analysis of population-based studies, possibly because of the inclusion of children and adolescents with comorbid autism spectrum disorder and cerebral palsy secondary to neonatal encephalopathy in the GBD.^32,33  South Asia substantially contributed to the reported global estimates for intellectual disability, and the age-specific GBD estimates for this region are supported by a recent robust population-based study from India in which a prevalence of 5.2% was reported.³⁴  Sensory disabilities are perhaps the most researched childhood disabilities worldwide. Several studies in school-aged children support the reported estimates for vision loss and hearing loss, with most suggesting that the GBD estimates were conservative.^35–38  For example, in 1 systematic review of childhood hearing loss, a prevalence of between 0.8% and 46.7% was reported across 26 well-conducted studies from different regions,³⁶  whereas 2 population-based studies from Canada and the United Sates revealed estimates of between 4.5% and 7.9%, compared with a GBD estimate of 2.2% for North America. More crucially, these findings underscore the need to develop local capacity toward early identification and timely on-going support for children and adolescents with any disability and their families, especially in high-burden LMIC. The unique challenges faced by children with disabilities as they transition into adolescence also need to be recognized and addressed.³⁹ 

This study has several limitations worth emphasizing. Most importantly, our main finding must be regarded as a conservative estimate of children with disabilities in general because this study was restricted to 4 conditions. The inclusion of conditions such as cerebral palsy without comorbid intellectual disability would have increased the reported estimates in our article. Additionally, the GBD 2017 has methodologic limitations that have been extensively described in previous publications in accordance with the Guidelines for Accurate and Transparent Health Estimates Reporting.^16–18  For example, the 95% UIs around estimates for regions with sparse data are still wide. Most of the uncertainty in the YLD estimates results from the current limitations in the determination of disability weights that may be minimized in the future by removing some of the ambiguities in lay descriptions and increasing the volume of survey data. In addition, despite the continuous efforts toward improving the GBD methodology, concerns remain on estimating the prevalence of disabilities solely as sequelae of health conditions.^2,3,40  Furthermore, the estimates do not fully reflect the complex and dynamic relationship between health conditions and contextual personal or environmental factors, as envisaged under the ICF, and so provide a limited picture of disability.¹⁵  Finally, it was difficult to completely and precisely account for children with multiple disabilities and across multiple developmental domains. These considerations reveal the need for complementary nationally representative disability data, such as those published periodically by the United Nations Children’s Fund and the WHO, that can be used as additional data inputs for the GBD. Although estimates of disabilities from both statistical modeling and household surveys are not definitive, they reflect current knowledge and the best available evidence to inform policies and interventions.

In the most recent GBD, it is shown that the number of children and adolescents with disabilities globally far exceeds the estimate in 2004. The burden of disability is substantial from early childhood and corroborates our earlier report on the need to address the quality of life of many beneficiaries of the child survival programs during the era of the Millennium Development Goals (2000–2015).¹⁸  The risk likely to be faced by the affected children and adolescents of not realizing their full potential, especially in LMIC, as envisaged under the SDGs era, is real and disturbing. Regardless of the inherent limitations of modeled estimates, the findings from our analysis are plausible and insightful. These findings shed light on the high and growing health needs among child survivors that warrant significant investments and should be a wake-up call to public health leaders and advocates globally. A committed global leadership will ensure that these and other vulnerable children and adolescents are truly not left behind in accordance with the obligations under the Convention on the Rights of the Child and the Convention on the Rights of Persons with Disabilities.

We thank Christopher J.L. Murray, Theo Vos, and Mohsen Naghavi of the Institute for Health Metrics and Evaluation for their overall guidance of the substantive GBD study that formed the basis of this analysis.

GBD

Global Burden of Disease Study

ICD-9

International Classification of Disease, Ninth Revision

ICD-10

International Classification of Disease, 10th Revision

ICF

International Classification of Functioning, Disability, and Health

LMIC

low- and middle-income countries

SDG

sustainable development goal

UI

uncertainty interval

WHO

World Health Organization

YLD

years lived with disability