According to the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM-5), autism spectrum disorder (ASD) is a condition characterised by persistent deficits in social communication and interaction and restricted, repetitive patterns of behaviour, interests or activities. Symptoms must be present in the early developmental period, cause clinically significant impairment and not be better explained by an intellectual disability or global developmental delay. 1 ASD affects approximately 1% of children worldwide, imposing significant health, social and economic burdens on affected individuals, their families and communities. Reference Issac, Halemani, Shetty, Thimmappa, Vijay and Koni2,Reference Salari, Rasoulpoor, Rasoulpoor, Shohaimi, Jafarpour and Abdoli3 Globally, ASD prevalence, incidence and disability-adjusted life-years (DALYs) increased by nearly 40% between 1990 and 2019. Reference Li, Yang, Chen, Fang, Zhang and Yin4 Although high Socio-Demographic Index (SDI) regions exhibit higher age-standardised rates, the low rates observed in many low- and middle-SDI regions, particularly in Africa, reflect widespread under-detection and underreporting resulting from limited diagnostic infrastructure and awareness. Reference Aderinto, Olatunji and Idowu5,Reference Li, Chen, Li, Gu, Xia and Gong6
Although ASD is widely recognised as a global public health concern, itsburden in Ethiopia remains poorly characterised owing to limited epidemiological data and diagnostic capacity. Prevalence estimates extrapolated from global findings suggest that more than half a million Ethiopian children may be living with ASD or related developmental disorders; however, awareness, diagnostic capacity and support services remain limited. Reference Alemnh7,Reference Tsegaye, Menbere, Godie and Birhanu8 Stigma, limited professional knowledge and inadequate healthcare facilities hinder early detection and intervention, contributing to social exclusion and long-term disability. Reference Kantawala, Abu-Bakr, Kasini, Ndayambaje, Ian Soh and Nazir9,Reference Turnock, Langley and Jones10 These challenges are further shaped by unique sociocultural factors. Qualitative studies indicate that neurodevelopmental differences are often attributed to spiritual causes or divine will, delaying diagnosis and contributing to social exclusion. Reference Asmare, Taye, Kotecho, Mishna and Regehr11,Reference Zeleke, Asfaw, Lee, King and Long12 While such research has captured caregiver experiences and societal challenges, an important gap remains: the national disease burden and future trends remain unquantified.
This study aims to fill a critical gap in knowledge regarding the trends of prevalence, incidence and DALYs associated with ASD in Ethiopia among children and adolescents using data from the Global Burden of Disease data between 1990 and 2023, with projections to 2050. Quantifying these epidemiological measures will provide evidence to inform policymakers, healthcare professionals and advocacy organisations in designing targeted interventions and strengthening the health system. Understanding trends in the burden of ASD will facilitate planning for diagnostic and rehabilitative services tailored to Ethiopia’s sociocultural dynamics. Furthermore, this study may help raise awareness of this long-neglected area of child and adolescent mental health, reduce the stigma surrounding it and encourage increased investment in resources. The findings will contribute valuable insights to the global literature on ASD in low-resource settings and support progress towards universal health coverage for people with developmental disabilities in Ethiopia.
Method
Study design and data source
This study performed a secondary analysis of estimates from the GBD 2023 study to assess the burden of ASD among children and adolescents aged 0–19 years in Ethiopia from 1990 to 2023. The GBD provides standardised location–year–sex–age metrics on health loss, accessible through the Institute for Health Metrics and Evaluation (IHME) Global Health Data Exchange (GHDx) Results Tool (http://ghdx.healthdata.org/gbd-results-tool). The GBD 2023 iteration benefits from methodological refinements that improve estimation precision in data-sparse settings such as Ethiopia through enhanced regional borrowing and covariate modelling.
Case definition and estimation
ASD in the GBD 2023 study was defined according to ICD-10 codes F84.0–F84.9. Studies using alternative diagnostic criteria, including DSM-based definitions, were harmonised through statistical crosswalk adjustments to ensure comparability across data sources.
GBD draws on multiple data sources, including population-based studies, surveys, published literature, population censuses, household surveys, demographic surveillance systems, disease registries, health service utilisation data and disease notifications. These primary data are generated by clinicians, trained health workers, researchers and survey teams, with diagnoses based on standardised frameworks such as ICD and DSM criteria, although the specific diagnostic instruments and screening tools may vary across settings. These differences inevitably introduce measurement variability into the primary data despite the harmonisation procedures applied within the GBD framework. To account for this variability, GBD produces internally consistent epidemiological estimates using DisMod-MR 2.1, a Bayesian meta-regression modelling framework. In settings with limited primary epidemiological data, such as Ethiopia, the model uses hierarchical Bayesian methods and incorporates regional information and covariates, including the Socio-Demographic Index (SDI), to generate age-, sex-, location- and year-specific estimates. Studies using earlier diagnostic classifications are adjusted to align with the ICD-10 definition, ensuring comparability across the 1990–2023 time series. The study population included Ethiopian children and adolescents aged 0–19 years, with analyses stratified by sex and 5-year age groups.
Outcome measures
The primary epidemiological measures evaluated in this study were prevalence, incidence and DALYs. As a non-fatal condition in the GBD framework, ASD has a Years of Life Lost (YLL) component of zero. Consequently, the DALY estimate is numerically identical to the Years Lived with Disability (YLD) estimate. For this analysis, we therefore used DALYs as the comprehensive measure of non-fatal health loss attributable to ASD in the Ethiopian paediatric population. All rates are reported per 100 000 population and are accompanied by 95% uncertainty intervals.
Temporal trend analysis
Temporal trends in ASD prevalence, incidence and DALYs from 1990 to 2023 were assessed using segmented (joinpoint) regression analysis. This method identifies statistically significant changes in trend slopes (joinpoints) and partitions the study period into distinct linear segments on the log scale. For each identified segment, the annual per cent change (APC) was estimated using a log-linear regression model and calculated as:
${\rm{APC}} = \;({e^{\beta i}}\; - \;1)\; \times \;100\%,$
where βi is the slope coefficient for segment i.
For age-standardised prevalence, incidence and DALY rates, the average annual per cent change (AAPC) was used to summarise the long-term trend over the entire study period (1990–2023). The AAPC was calculated as the geometric mean annual change between the start and end of the period using the formula:
${\rm{AAPC}} = \left[ {{{\left( {{{R\;\rm{end}} \over {R\;\rm{start}\;}}} \right)}^{1/n}} - 1} \right] \times \;100\%,$
where R start and R end denote the age-standardised rate at the beginning and end of the study period, respectively, and n represents the number of years between the two time points (n = 33).
Sex-specific and age-specific analyses
Analyses were stratified by sex to enable comparison of ASD burden between males and females, with separate estimates produced for each group to highlight disparities. Additionally, age-specific DALY rates were estimated for different age subgroups within the under-20 population to identify which cohorts bear the greatest disability burden.
Projection of future burden
To estimate future ASD burden, Autoregressive Integrated Moving Average (ARIMA) time series models were fitted to historical ASD incidence and DALYs data from 1990 to 2023. Model selection for ARIMA parameters (p, d, q) was determined by minimising Akaike’s Information Criterion (AIC) and assessment of residual diagnostics to ensure goodness of fit.
The fitted ARIMA models were used to forecast incidence and DALYs to 2050, with 95% uncertainty intervals reflecting forecast variability. These projections provide a quantitative evidence base to inform long-term health system planning and resource allocation for ASD-related services in Ethiopia.
Software
All statistical analyses, modelling and data visualisation were performed using Python (version 3.12) within the Google Colaboratory environment. Segmented joinpoint regression parameter estimation and log-linear modelling were processed using data structures within the pandas and NumPy scientific computing libraries. Autoregressive time series components and predictive ARIMA modelling were implemented using the statsmodels library. All publication-grade figures, trend lines and shaded 95% uncertainty intervals were generated using the matplotlib package.
Results
National prevalence and incidence trends of ASD among individuals aged under 20 years
By 2023, the prevalence of ASD among individuals aged under 20 years in Ethiopia had increased to 693 644 cases (95% uncertainty interval: 243 353–1 951 468), from 271 805 (95% uncertainty interval: 117 722–649 105) in 1990. This represents a 155% increase over the past three decades. A continuous segmented joinpoint regression analysis identified three distinct temporal phases in ASD prevalence among children and adolescents in Ethiopia from 1990 to 2023. During the initial period (1990–1996), prevalence increased at an APC of 4.48%. Following the first joinpoint in 1996, prevalence continued to increase during a second phase (1996–2011) at an APC of 2.78%. By 2010, the fitted prevalence had reached 522 557 cases. The final period (2011–2023) was characterised by continued but slower growth, with an APC of 2.20%. By 2023, the fitted model estimated a prevalence of 697 001 cases (95% uncertainty interval: 693 826–700 190).
In 2023, the annual incidence of ASD in Ethiopia reached 47 274 cases (95% uncertainty interval: 10 492–139 127), increasing from 24 576 cases (95% uncertainty interval: 11 049–52 840) in 1990. This represents a 92.3% increase in annual incident cases over the study period. Joinpoint regression analysis identified three distinct phases of growth in the incidence trend. From 1990 to 2000, annual incidence increased at an APC of 3.22%. During this period, the fitted number of new cases increased from 25 037 to 34 377. From 2000 to 2014, the growth rate underwent a significant deceleration. The APC decreased to 1.19%, indicating a period of more gradual increase in the annual number of new ASD cases. In the final decade, the incidence rate increased significantly, with an APC of 1.73%, with the fitted incidence reaching 47 315 by 2023 (Table 1).
Prevalence and incidence trends of ASD among individuals aged under 20 years in Ethiopia
Table 1 Long description
The table presents data on the prevalence and incidence trends of ASD among individuals aged under 20 years in Ethiopia. It is divided into two main sections: prevalence trend and incidence trend. Each section contains data for specific years, observed prevalence or incidence with 95% uncertainty intervals, fitted prevalence or incidence with 95% uncertainty intervals, trend periods, and annual percent change (APC). The table has 13 rows and 5 columns. Column headers are Year, Observed prevalence (95% UI), Fitted prevalence (95% UI), Trend period, and APC (%). Row labels include specific years from 1990 to 2023. The prevalence trend section shows data for the years 1990, 1995, 2000, 2005, 2010, 2015, 2020, and 2023, with corresponding observed and fitted prevalence values, trend periods, and APC percentages. The incidence trend section shows data for the same years, with corresponding observed and fitted incidence values, trend periods, and APC percentages. Notable trends include an overall increase in both prevalence and incidence over the years, with specific periods of growth and deceleration.
ASD, autism spectrum disorder; UI, uncertainty interval; APC, Annual Percent Change.
National prevalence trends of ASD among individuals aged under 20 years by sex
In 2023, the prevalence of ASD among males aged under 20 years in Ethiopia was estimated at 508 314 cases (95% uncertainty interval: 178 960–1 423 690), representing a 155.45% increase from 1990, when prevalence was 198 990 (95% uncertainty interval: 86 525–474 889). From 1990 to approximately 1996, prevalence increased at an APC of 4.49%. Between 1996 and 2012, prevalence continued to increase at an APC of 2.77%, and from 2012 to 2023, it increased at an APC of 2.15%. Similarly, among females, the prevalence reached 185 330 cases (95% uncertainty interval: 64 393–527 778) in 2023, representing a 154.52% increase from 72 815 cases (95% uncertainty interval: 31 197–175 363) in 1990. Between 1990 and 1996, female prevalence increased at an APC of 4.47% per year. Following the trend in change in 1996, the rate of increase slowed to an APC of 2.79% annually to 2010, before increasing at an APC of 2.23% per year from 2010 to 2023 (Fig. 1).
Prevalence of ASD in Ethiopians aged under 20 years, by sex (1990–2023). (a) males; (b) females. ASD, autism spectrum disorder; APC, annual per cent change.
National incidence trends of ASD among individuals aged under 20 years by sex (1990–2023)
In 1990, the number of incidence ASD cases among males was 18,063 (95% uncertainty interval: 8131–38 697), increasing to 34 796 (95% uncertainty interval: 7755–102 200) in 2023, representing a 92.64% increase. Male incidence increased at an APC of 3.23% per year between 1990 and 2000. Followingthe change in trend in 2000, the rate of increase slowed to an APC of 1.19% per year between 2000 and 2014. From 2014 to 2023, the incidence continued to increase at an APC of 1.73% per year. Similarly, among females, the number of ASD cases in 1990 was 6514 (95% uncertainty interval: 2892–14 144) and increased to 12 478 cases (95% uncertainty interval: 2737–36 927) in 2023, an increase of 91.56%. Female incidence increased from 1990 to 2000 by an APC of 3.19% per year. Following the change in trend in 2000, the rate of increase slowed to an APC of 1.17% per year until 2014. From 2014 to 2023, the incidence increased at an APC of 1.73% per year (Fig. 2).
Incidence trend of ASD in Ethiopians aged under 20 years, by sex (1990–2023). (a) males; (b) females. ASD, autism spectrum disorder; APC, annual per cent change.
Fig. 2 Long description
Two line graphs depict the incidence trend of autism spectrum disorder in Ethiopians aged under 20 years, separated by sex, from 1990 to 2023. Panel A shows the trend for males, while Panel B shows the trend for females. Each graph includes a fitted trend line, observed incidence points, and 95% uncertainty intervals. The x-axis represents the year, ranging from 1990 to 2023, and the y-axis represents the number of new cases. The graphs are divided into three segments with different annual percent changes (APC) indicated: 1990-2000, 2000-2014, and 2014-2023. In Panel A, the APC for males is 3.23% from 1990 to 2000, 1.19% from 2000 to 2014, and 1.73% from 2014 to 2023. In Panel B, the APC for females is 3.19% from 1990 to 2000, 1.17% from 2000 to 2014, and 1.73% from 2014 to 2023. The observed incidence points generally follow the fitted trend lines, with some variability within the uncertainty intervals. The number of new cases increases over time in both graphs, indicating a rising trend in ASD incidence for both males and females.
National trends in DALYs attributable to ASD among individuals aged under 20 years (1990–2023)
In 1990, the number of DALYs attributable to ASD among Ethiopians aged under 20 years was 51 321 (95% uncertainty interval: 20 116–117 231), increasing to 134 386 (95% uncertainty interval: 39 898–400 296) in 2023. This represents a 161.85% increase over the study period. The number of DALYs increased between 1990 and 1996 at an APC of 4.53%. Following the change in trend in 1996, the rate of increase slowed to an APC of 2.84% per year from 1996 to 2013. From 2013 to 2023, the rate of increase slowed further, with an APC of 2.19%.
Among males, the number of DALYs attributable to ASD in 1990 was 37 577 (95% uncertainty interval: 14 890–85 507). This increased to 98 471 (95% uncertainty interval: 29 532–291 771) in 2023, an increase of 162.05%. Similar to the overall trend, DALYs among males increased between 1990 and 1996 at an APC of 4.51% per year. Following the change in trend in 1996, the rate of increase slowed to an APC of 2.85% per year until 2013. From 2013 to 2023, the rate of increase slowed further, with an APC of 2.19% per year. Among females, the number of DALYs attributable to ASD was 13 744 (5240–32 082) in 1990 and increased to 35 915 (95% uncertainty interval: 10 366–108 525) in 2023, representing an increase of 161.31%. DALYs among females increased between 1990 and 1996 at an APC of 4.55% per year. Following the change in trend in 1996, the rate of increase slowed to an APC of 2.81% per year until 2013. From 2013 to 2023, DALYs continued to increase at an APC of 2.20% per year (Table 2).
Trends in ASD DALYs among individuals aged under 20 years in Ethiopia, 1990-2023
Table 2 Long description
The table presents trends in ASD DALYs among individuals aged under 20 years in Ethiopia from 1990 to 2023. It is divided into three sections: trends for both sexes, trends for males, and trends for females. Each section includes columns for the year, observed DALYs with 95% uncertainty intervals, fitted DALYs with 95% uncertainty intervals, trend periods, and the annual percent change (APC). The table has 18 rows and 5 columns. Row 1: Year, Observed DALY (95% UI), Fitted DALY (95% UI), Trend period, APC (%). Row 2: Trends in ASD DALYs for both sexes (1990-2023). Row 3: 1990, 51321 (20116-117231), 51640 (51339-51943), 1990-1996, 4.53. Row 4: 1995, 64124 (27665-133048), 64431 (64244-64619), 1996-2013, 2.84. Row 5: 2000, 75492 (32660-158225), 75340 (7415-75533), 2013-2023, 2.19. Row 6: 2005, 86900 (38124-178806), 86678 (86509-86848). Row 7: 2010, 99741 (4071-222716), 99723 (99463-99983). Row 8: 2015, 113183 (39603-285792), 113278 (112987-113570). Row 9: 2020, 126424 (40608-356259), 126240 (125833-126647). Row 10: 2023, 134386 (39898-400296), 134718 (134070-135370). Row 11: Trends in ASD DALYs for males (1990-2023). Row 12: 1990, 37577 (14890-85507), 37831 (37601-38062), 1990-1996, 4.51. Row 13: 1995, 46959 (20347-97431), 47177 (47034-47320), 1996-2013, 2.85. Row 14: 2000, 55273 (24114-116045), 55181 (55034-55329), 2013-2023, 2.19. Row 15: 2005, 63681 (28368-130655), 63518 (63388-63648). Row 16: 2010, 73123 (29993-162943), 73114 (72916-73313). Row 17: 2015, 83021 (29256-208718), 83070 (82848-83293). Row 18: 2020, 92717 (29764-260977), 92553 (92243-92865). Row 19: 2023, 98471 (29532-291771), 98755 (98259-99253). Row 20: Trends in ASD DALYs for females (1990-2023). Row 21: 1990, 13744 (5240-32082), 13810 (13734-13886), 1990-1996, 4.55. Row 22: 1995, 17164 (7318-35617), 17254 (17207-17302), 1996-2013, 2.81. Row 23: 2000, 20219 (8622-42324), 20159 (20110-20208), 2013-2023, 2.20. Row 24: 2005, 23220 (9897-48276), 23160 (23117-23203). Row 25: 2010, 26619 (10378-59339), 26608 (26543-26674). Row 26: 2015, 30162 (10352-77075), 30208 (30134-30281). Row 27: 2020, 33707 (10683-95282), 33686 (33584-33789). Row 28: 2023, 35915 (10366-108525), 35963 (35799-36128).
ASD, autism spectrum disorder; DALY, disability-adjusted life-year; UI, uncertainty interval; APC, Annual Percent Change.
Age-standardised prevalence, incidence and DALY trajectories of ASD among Ethiopians aged under 20 years, 1990–2023
The national age-standardised ASD prevalence rate per 100 000 population (95% uncertainty interval was 1185 (416–3333) in 2023, representing an increase from 950 (412–2269) in 1990, with an AAPC of 0.67%. Across all analysed age cohorts, a consistent but modest increase in the age-standardised prevalence rate was observed between 1990 and 2023. Children aged under 5 years consistently had the highest age-standardised prevalence rate, increasing from 985 (95% uncertainty interval: 437–2107) in 1990 to 1269 (95% uncertainty interval: 332–3689) in 2023. The age-standardised DALY rate per 100 000 population was 230 (95% uncertainty interval: 68–684) in 2023, increasing from 179 (95% uncertainty interval: 70–410) in 1990, with an AAPC of 0.76%. Similarly, children aged under 5 years and late adolescents consistently had the highest and lowest DALY rates, respectively. Conversely, the age-standardises incidence rate declined slightly, from 86 (95% uncertainty interval: 39–185) per 100 000 in 1990 to 81 (95% uncertainty interval: 18–238) in 2023, with an AAPC of −0.20%. Trends in age-standardised prevalence, incidence and DALY rates between 1990 and 2023 are shown in Fig. 3.
Trends in age-standardised indicators among individuals aged under 20 years in Ethiopia (1990–2023). (a) age-standardised prevalence rate; (b) age-standardised DALY rate; (c) age-standardised incidence rate. ASD, autism spectrum disorder; DALY, disability-adjusted life-year.
Fig. 3 Long description
Three line graphs depict trends in autism spectrum disorder (ASD) age-standardized indicators among individuals aged under 20 years in Ethiopia from 1990 to 2023. Panel A shows the age-standardized prevalence rate, Panel B shows the age-standardized DALY rate, and Panel C shows the age-standardized incidence rate. Each graph includes multiple lines representing different age groups: under 5 years, 5-9 years, 10-14 years, 15-19 years, and 20 years and older. The x-axis represents the year, ranging from 1990 to 2023, and the y-axis represents the rate per 100,000. Panel A’s y-axis ranges from 900 to 1250, Panel B’s y-axis ranges from 170 to 250, and Panel C’s y-axis ranges from 0 to 250. Each graph shows an upward trend in the rates over the years, with the prevalence and DALY rates increasing steadily across all age groups. The incidence rate remains relatively stable over the years. The lines are color-coded and labeled accordingly for each age group. The graphs highlight the increasing burden of ASD in Ethiopia over the specified period.
Projected ASD burden among individuals aged under 20 years in Ethiopia to 2050
ARIMA modelling projected a steady increase in the incidence of ASD among individuals aged under 20 years in Ethiopia. Starting from a 2023 baseline of approximately 81 per 100 000, the incidence rate is projected to increase to approximately 86 per 100 000 by 2050, representing an overall increase of 6%. The projection is characterised by a widening 95% uncertainty interval over the forecast period, reflecting increasing uncertainty in the long-term projections. By 2050, the projected incidence rate is expected to range from 70 to 101 per 100 000.
The age-standardised DALY rate attributable to ASD is likewise projected to increase steadily. From a 2023 baseline of approximately 230 DALYs per 100 000, the rate is projected to increase to approximately 280 DALYs per 100 000 by 2050. This shift reflects a total percentage increase of 21.73% in the disease burden over the 27-year period. The 95% uncertainty interval expands progressively over the projection period, reflecting increasing uncertainty in the long-term forecasts. By 2050, the projected DALY rate is expected to range from 255 to 310 per 100 000 (Fig. 4).
Projected incidence and DALY rate for ASD in Ethiopians aged under 20 years (2023–2050). (a) incidence; (b) DALYs. ASD, autism spectrum disorder; UI, uncertainty interval; DALY, disability-adjusted life-year.
Fig. 4 Long description
Two line graphs depict projections for autism spectrum disorder (ASD) in Ethiopia from 2023 to 2050. Panel A shows the ASD incidence rate projection, while Panel B shows the ASD disability-adjusted life-years (DALY) rate projection. Panel A: The line graph displays the historical incidence rate in blue, the projected incidence rate in red dashed line, and the 95 percent uncertainty interval (UI) for the forecast in pink shading. The x-axis represents the year from 1990 to 2050, and the y-axis represents the incidence rate per 100,000. The historical data shows a slight decline until around 2015, after which the projected data indicates a gradual increase in incidence rates. Panel B: The line graph shows the historical DALY rate in green, the projected DALY rate in red dashed line, and the 95 percent UI for the forecast in pink shading. The x-axis represents the year from 1990 to 2050, and the y-axis represents the DALY rate per 100,000. The historical data shows a steady increase in DALY rates, and the projected data continues this upward trend, with the rates rising more steeply after 2025.
Discussion
The present study describes the changing burden of ASD among Ethiopian children and adolescents over the 33-year period from 1990 to 2023, with projections to 2050. The longitudinal analysis, based on the Global Burden of Disease dataset, delineates crucial epidemiological transitions in prevalence, incidence and DALYs, each accompanied by requisite interpretations and implications for public health.
The prevalence of ASD increased by 155%, from 271 805 cases in 1990 to over 693 644 in 2023, highlighting the growing burden of neurodevelopmental disorders among Ethiopian children and adolescents. Joinpoint analysis identified three distinct phases (1990–1996, 1996–2011 and 2011–2023), with the APC declining from 4.48% to 2.78% and then to 2.20%, indicating that prevalence continued to increase but at a progressively slower rate. Similarly, the national age-standardised prevalence rate increased from 950 per 100 000 in 1990 to 1185 per 100 000 in 2023. This indicates that the burden is expanding even when adjusting for changing population demographics. Children under 5 years consistently had the highest age-standardised prevalence, reaching 1269 per 100 000 in 2023.
The increasing prevalence of ASD in Ethiopia over the past three decades likely reflects a combination of improved diagnostic detection and a genuine increase in the underlying neurodevelopmental burden. Since the early 2000s, Ethiopia has implemented gradual health system reforms, including the expansion of mental health services through the National Mental Health Strategy and its subsequent iterations. 13,14 Although child psychiatrists remain scarce, specialist training programmes for mental health professionals, general practitioners and health officers have increasingly incorporated developmental screening. In recent years, heightened advocacy by parent associations and local non-governmental organisations (NGOs) has raised public and professional awareness, potentially reducing diagnostic delays. Community-based rehabilitation programmes and school health initiatives may also have improved the identification of children with developmental disorders. However, these improvements remain unevenly distributed. The persistent accumulation of cases in the most recent period, despite a lower APC, suggests that both improvements in case detection and changes in underlying risk factors, including environmental exposures associated with rapid urbanisation, perinatal complications and genetic susceptibility, have contributed to the observed long-term trend. Reference Issac, Halemani, Shetty, Thimmappa, Vijay and Koni2,Reference Talantseva, Romanova, Shurdova, Dolgorukova, Sologub and Titova15
Similarly, the incidence of ASD increased by 92.3%, from 24 576 cases in 1990 to 47 274 in 2023. While absolute numbers surged, the age-standardised incidence rate declined slightly from 86 per 100 000 in 1990 to 81 per 100 000 in 2023. This divergence suggests that while the annual number of incident cases is increasing due to demographic shifts, the underlying age-standardised risk has remained relatively stable. The trend followed a fluctuating pattern, with an APC of 3.22% from 1990 to 2000, slowing to 1.19% between 2000 and 2014 before increasing to 1.73% per year from 2014 to 2023.
The deceleration in APC between 2000 and 2014 may reflect limitations in diagnostic capacity rather than a true stabilisation of ASD risk. During this period, Ethiopia’s mental health system remained substantially under-resourced, with few specialised diagnostic centres and limited training in neurodevelopmental disorders among frontline healthcare workers, constraining case ascertainment. The subsequent increase after 2014 aligns with greater policy attention to developmental disabilities through Ethiopia’s Health Sector Transformation Plans and the gradual introduction of initiatives such as community-based rehabilitation programmes. These efforts have strengthened, although not fully addressed healthcare infrastructure and may have improved identification of previously overlooked cases. 13,14
This pattern, consistent with the prevalence trend, suggests that increases in recorded ASD incidence are likely driven partly by improvements in detection, although diagnostic capacity remains incomplete. Similar patterns have been reported in other low-SDI countries, where health system strengthening has been associated with increases in recorded ASD prevalence, highlighting the important contribution of improved detection to observed epidemiological trends. Reference Franz, Chambers, von Isenburg and de Vries16 These findings underscore the urgent need for sustained investment in diagnostic tools, specialised training and expanded screening programmes to facilitate early identification and timely intervention.
DALYs attributable to ASD increased by 162%, from 51 321 in 1990 to 134 386 in 2023, a trend parallel to that for prevalence and incidence. The age-standardised DALY rate also increased from 179 to 230 per 100 000, indicating that the disability burden attributable to ASD has increased. The highest DALY rates were consistently observed among children aged under 5 years, highlighting the substantial impact of ASD-related disability during early childhood.
Critically, the observed increase in ASD burden must be interpreted within the context of shifting diagnostic sensitivity. As awareness and screening capacity improve in Ethiopia, there is an increased probability of detecting milder cases on the spectrum that were previously overlooked. Notably, the age-standardised DALY rate increased by 28.5% (increasing from 179 to 230) between 1990 and 2023. This trend suggests that the increasing burden cannot be attributed solely to improved detection of milder cases. Rather, it indicates an overall increase in population-level disability associated with ASD. For Ethiopia’s health system, these findings highlight the need for a tiered service delivery model that prioritises intensive rehabilitative resources for high-severity cases while using community-based screening and support for those with milder presentations.
Sex-specific comparison of ASD prevalence, incidence and DALYs
Sex-stratified analyses showed that males consistently had a substantially higher ASD prevalence than females throughout the study period. In 2023, the prevalence of ASD was 508 314 cases among males compared with 185 330 among females. This indicates that male prevalence was approximately 2.74 times higher than female prevalence in 2023, a ratio that has widened since 1990. Similar results regarding male predominance have been reported in global studies Reference Salari, Rasoulpoor, Rasoulpoor, Shohaimi, Jafarpour and Abdoli3,Reference Santomauro, Erskine, Herrera, Miller, Shadid and Hagins17 although contrasting findings have been described in Southeast Asia. Reference Liu, Gong, Ou and Chen18 Although prevalence increased in both sexes over the study period, the rate of increase was marginally higher among males.
Similarly, males consistently exhibited higher ASD incidence rates than females. In 2023, male incidence reached 34 796 cases compared with 12 478 for females, maintaining a ratio where males are approximately 2.79 times more likely to be diagnosed annually, consistent with global estimates. Reference Loomes, Hull and Mandy19 Males exhibited slightly higher growth rates across trend segments, potentially reflecting earlier diagnosis resulting from more overt externalising behaviours, differences in healthcare-seeking behaviour or biological susceptibility. Reference Li, Liang, Miao, Li and CJFiP20,Reference Horwitz, Vos, De Bildt, Greaves-Lord, Rommelse and Schoevers21
Sex differences were also evident for DALYs, with males consistently experiencing a greater disability burden than females throughout the study period. In 2023, the number of DALYs attributable to ASD was 98 471 among males compared with 35 915 among females, corresponding to a male-to-female ratio of 2.74:1, aligning with global patterns. Reference Li, Chen, Li, Gu, Xia and Gong6,Reference Santomauro, Erskine, Herrera, Miller, Shadid and Hagins17,Reference Mao, Cheng, Zhang and Zhang22
However, this observed male predominance should be interpreted with caution in the Ethiopian context. ASD in females is frequently underdiagnosed because of more subtle or internalising clinical presentations, including social camouflaging behaviours that may mask core symptoms. This challenge may be amplified in low-resource settings such as Ethiopia, where limited awareness, cultural perceptions and gender-related barriers to healthcare access may further reduce the identification of affected girls. Consequently, the burden of ASD among females may be underestimated in GBD-based estimates. This persistent disparity, whether reflecting true biological differences or diagnostic bias, underscores the need for sex-specific screening protocols and community awareness strategies to ensure equitable early detection and intervention for all children.
Age-standardised DALY rates peak in children aged under 5 years, representing a critical intervention window where integrating ASD screening into primary care and early child development programmes could substantially reduce long-term disability. Health systems must scale behavioural, educational and social support services to address this escalating need.
The ARIMA model projection shows a continued upward trend in ASD burden through to 2050, with the incidence rate expected to reach 86 per 100 000 and the DALY rate 280 per 100 000. This corresponds to a 22% increase in the DALY rate over the next 27 years. A consistent upward trend has been reported in global studies. Reference Li, Liang, Miao, Li and CJFiP20,Reference Zhu, Zhang, Wang and Yang23 However, the trajectory of ASD burden will markedly differ with future advances in diagnosis, shifts in socio-environmental factors and healthcare investments. These forecasts highlight the need for flexible policy frameworks and robust epidemiological surveillance systems that are capable of responding to unpredictable changes in the disease burden over time.
This study has several strengths, including the use of robust GBD 2023 data and rigorous statistical methods, including joinpoint regression and ARIMA modelling, to define temporal trends and long-term projections in the Ethiopian paediatric population. However, limitations remain. The study relies on GBD-modelled secondary data, which uses regional covariates including the SDI to compensate for the limited availability of primary epidemiological data in Ethiopia. Consequently, the slight decline in age-standardised incidence rate may reflect modelling assumptions rather than a true biological trend, and the overall burden may be underestimated in healthcare-limited settings where case ascertainment is low. Diagnostic overshadowing, where ASD symptoms are misattributed to intellectual disability, may further contribute to under-detection of milder cases. In addition, although GBD study applies crosswalk methods to harmonise diagnostic criteria across time, these adjustments may not fully capture temporal changes in diagnostic practice, awareness and access to diagnostic services. Consequently, part of the observed increase in ASD prevalence and DALYs may reflect improved case detection over time rather than a true increase in disease occurrence, and these temporal trends should therefore be interpreted with caution. Finally, the absence of primary epidemiological data limits exploration of localised risk factors.
Implications for policy and practice
The substantial increase in Ethiopia’s ASD metrics over the past three decades culminated in a profound public health reality, with approximately 700 000 children and adolescents affected by the condition in 2023. Between 1990 and 2023, the prevalence and incidence of ASD increased by 155% and 92%, respectively. This escalating trend is closely mirrored by a 162% increase in DALYs, reflecting a heavy and growing disability burden. Moreover, the increases in the age-standardised prevalence and DALY rates indicate that this growing burden cannot be explained solely by population growth or demographic change.
Furthermore, the analysis revealed persistent sex differences, with males consistently experiencing a higher burden than females across prevalence, incidence and DALYs. These historical trends, paired with projections indicating that the national burden will remain high through 2050, demonstrate that ASD represents a critical, long-term public health priority.
A comprehensive national strategy is essential to address the growing burden of ASD in Ethiopia. This strategy should prioritise strengthening epidemiological surveillance and diagnostic capacity by establishing a robust tracking system for ASD indicators, primarily through the integration of ASD screening into the existing Health Extension Program. Furthermore, targeted interventions and health system integration are critical; ASD care including behavioural, educational and social support services should be incorporated into primary healthcare to ensure accessibility. Finally, increasing public awareness through community-based programmes is fundamental to reducing stigma and improving the quality of life for children and adolescents with ASD.
Data availability
The data used in this study were obtained from the GBD 2023 database and are publicly available at https://ghdx.healthdata.org/gbd-2023.
Author contributions
M.A.A. contributed to conceptualisation, data curation, formal analysis, methodology, software, visualisation, writing of the original draft and review and editing of the manuscript. Y.A.D. contributed to formal analysis, methodology, writing of the original draft and review and editing of the manuscript. A.B.Z. contributed to data curation, formal analysis, methodology, writing of the original draft and review and editing of the manuscript. A.D.G. contributed to formal analysis, methodology, writing of the original draft and review and editing of the manuscript. B.M. contributed to data curation, formal analysis, methodology, writing of the original draft and review and editing of the manuscript. T.E.H. contributed to conceptualisation, formal analysis, methodology, software, writing of the original draft and review and editing of the manuscript. All authors have read and approved the final version of the manuscript for publication.
Funding
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Declaration of interest
None.
Ethical standards
Ethical approval was not required because this study involved secondary analysis of publicly available, anonymised data. All data were used in accordance with the Institute for Health Metrics and Evaluation (IHME) data use policies.
Open access
eLetters
No eLetters have been published for this article.