Global Disease Burden of Traumatic Joint Dislocation from 1990 to 2021 and its prediction to 2045

Abstract

Background

Traumatic joint dislocations of the hip, knee, and shoulder (DOH, DOK, and DOS) significantly impact global healthcare. This study assesses the global burden of joint dislocations using the Global Burden of Disease (GBD) 2021 database, focusing on their association with the socio-demographic index (SDI).

Methods

Data from the GBD 2021 are analyzed to determine the age-standardized rates (ASR) of incidence, prevalence, and years lived with disability (YLDs) for dislocations. We integrate the SDI with the concentration index, assessing disparities in the burden of these joint dislocations. Frontier analysis is performed to identify potential improvement areas and disparities among countries by development status. The age-period-cohort (APC) model projects the disease burden to 2045, with a focus on age and gender distributions and primary causes.

Results

From 1990 to 2021, the incidence, prevalence, and YLDs of DOH/DOK/DOS all increase, while ASRs decline, suggesting a deceleration in growth. YLDs of DOH, DOK, and DOS rise by 57.21%, 28.38%, and 15.48%, respectively. Men exhibit a higher burden, yet women show a steeper rise. Significant geographical variation exists, with lower SDI countries facing higher burdens. Falls and road injuries remain the main contributors to the burden, and lower-development countries demonstrate potential for reduction. Temporal trends vary by age, sex, and SDI, with projections indicating continued disparities to 2045.

Conclusions

Traumatic joint dislocations show marked heterogeneity in age, sex, and SDI, with the most significant differences in low-income regions. Research should prioritize policy development and targeted prevention and treatment strategies for groups at high-risk for joint dislocation to effectively mitigate the disease burden.

Huang et al. analyze the global burden of hip, knee, and shoulder dislocations using 1990–2021 data and project trends to 2045. Despite declining age-standardized rates, cases rise due to population growth, with men disproportionately affected, women cases increasing, and low-income countries bearing the highest burden.

Plain Language Summary

Hip, knee and shoulder dislocations cause significant pain and disability worldwide. Using global health data from 1990 to 2021, we analyzed how often these injuries occur, who is most affected, and how the burden may change by 2045. Although rates per capita have declined, the total number of cases continue to rise due to population growth. Men are more likely to experience dislocation, but cases among women are increasing faster. Low-income countries carry the greatest burden. These findings highlight the need for targeted prevention, especially in groups and regions at higher risk for joint dislocation, to reduce the impact of joint dislocations on individuals and health systems in the future.

Introduction

Joint dislocations occur often lead to trauma, causing pain, swelling, bruising, deformity, and immobility^1–3. Hip, knee, and shoulder are common sites of dislocation, frequently resulting from high-energy events such as motor vehicle accidents, sports injuries, and fall. Dislocations of the hip, knee and shoulder are more complex and difficult to treat than dislocations of the elbow and ankle, and are more likely to cause long-term complications^4–8. The National Electronic Injury Surveillance System (NEISS) reports that 3.6% of Americans experience sports-related dislocations annually^9,10.

Dislocations of the hip (DOH) typically involve severe posterior force on the knee with flexed hips, leading to leg shortening, adduction, and internal rotation¹¹. The main complication of DOH is damage to the femoral head’s vascular supply, potentially resulting in avascular necrosis¹². Dislocation of the knee (DOK) often result from severe trauma, leading to potential amputation and disability¹³. And DOK is a catastrophic injury that results in multiple ligament tears that can damage regional neurovascular structures¹⁴. Dislocation of shoulder (DOS) is primarily anterior, usually caused by excessive abduction and external rotation, which can cause nerve injuries and rotator cuff damage^3,15. Moreover, joint dislocation brings heavy economic burden to society. According to a 2011 study, the average direct medical cost for treating DOS is $612, for DOH it is $1103, and for DOK, the highest cost is $1888¹⁶.

Previous research on joint dislocations indicates that DOS account for approximately 38.7% and DO for 23.0% of cases¹⁷, while due to the lack of large-scale epidemiological studies on DOH, the global incidence, prevalence, years lived with disability (YLDs), and the proportion of joint dislocation are not clear. Due to the low incidence of injuries, studies on traumatic DOH are limited by small patient numbers, reducing their value^18–21. There is variation in reported incidence rates for traumatic DOH. The small sample size and single center led to significant differences in the results of previous studies^22–25. The recurrence rate of joint dislocation is high^26–29, so we need up-to-date data to keep an eye on the latest epidemiology of DOK and DOS. Given the significant burden of these injuries, it is crucial to investigate the global and regional epidemiology of DOH, DOK, and DOS, focusing on trends, risk factors, and healthcare burden.

In this study, we aim to assess the global burden, trends, and risk factors of hip, knee, and shoulder joint dislocations based on GBD 2021 data. This study will provide up-to-date epidemiological estimates and future projections for these three diseases with a view to providing a better understanding of disease burden patterns and characteristics.

Methods

Data source

The data for this study is obtained from the GBD 2021 database, available through the Institute for Health Metrics and Evaluation (IHME) GBD Results Tool³⁰. The 2021 GBD analysis covers 371 diseases and injuries across 204 countries and territories, and 811 localities from 1990 to 2021. The detailed methodology for GBD 2021 has been described elsewhere^31–33. The GBD database employs complex statistical methods to account for missing data and adjust for confounders, and its study design and methods have been extensively detailed in the existing GBD literature³¹. Briefly, raw data for GBD were collected from specific disease registries, health service contact data, vital registration systems, censuses, household surveys, and other sources. The data were then synthesized using Bayesian meta-regression modeling tools (dismod-2.1) and spatio-temporal Gaussian process regression (Detail in “Method” Appendix). The GBD study used deidentified data, and the University of Washington Institutional Review Board approved the waiver of informed consent. GBD 2021 follows the Guidelines for Accurate and Transparent Health Estimates in Population Health Research, ensuring methodological rigor and transparency³¹.

For non-fatal estimates, GBD utilized data sources including scientific literature, household surveys, epidemiological surveillance, disease registries, clinical informatics, and other sources. Cause-specific literature reviews encompassed online research databases, government/international organization websites, published reports, primary data, and datasets contributed by GBD Collaborators. Multiple data types were included to capture the broadest possible epidemiological information for each cause. Data were sourced from vital registration, verbal autopsy, registries, surveys, police records, and surveillance systems across all locations. As with every GBD release, we aimed to incorporate all available global data into our estimates. Data acquisition occurred not only through literature reviews, but also via proactive data seeking and collaborator network efforts to identify new datasets³¹.

This study extracted data on joint dislocations (hip, knee, and shoulder) due to natural injuries from the GBD 2021 database for 1990–2021. DOH was a painful event in which the ball joint of your hip comes out of its socket. It usually occurs from a significant traumatic injury ³⁴. DOK was defined as a loss of complete agreement between the articular surfaces of the distal femur and the proximal tibia³⁵. DOS involved separation of the scapular glenoid from the humerus at the glenohumeral joint³⁶. The codes for DOH, DOK and DOS in the International Classification of Diseases (ICD-11) are NC73.0, NC93.2, and NC13.0, respectively ³⁷.

We extracted estimates of the number and age-standardized rates for incidence, prevalence, and years lived with disability (YLDs), along with their corresponding 95% uncertainty intervals (UIs). Incidence refers to the number of new joint dislocations globally or regionally during the 1990–2021 period, while prevalence indicates the proportion of the population affected at specific points in time, such as in 1990 and 2021. YLDs quantify the healthy life years lost due to disabilities caused by hip, knee, or shoulder dislocations, reflecting the impact on quality of life. The UI represents the probability distribution around the true value, derived from 1000 sampling iterations, with the 95% UI calculated from the 2.5th and 97.5th percentiles³¹.

Additionally, this study employed the socio-demographic index (SDI), which assesses health-related socio-demographic development based on average income, education, and fertility rates^38,39. Each country or region was assigned an SDI value between 0 and 1, with higher values indicating a theoretically greater level of health-related development.

MR-BRT (meta-regression—Bayesian, regularized, trimmed)

To address biases in epidemiological data (e.g., differing case definitions or measurement methods), we applied correction factors estimated via MR-BRT (meta-regression—Bayesian, regularized, trimmed) network meta-regressions. MR-BRT utilizes linear and non-linear mixed effects models and fitting procedures. Correction factors were estimated using paired data points where two definitions/methods existed for the same age-sex-location-year⁴⁰. Example: past-year joint dislocation prevalence was adjusted down to point prevalence (the gold standard) using a factor informed by matched past-year and point prevalence data (applied via cause-specific MR-BRT corrections). Data missing sex reporting used cause-specific pooled within-study sex ratios. Data missing both age and sex underwent age-sex splitting. Sources reporting age and sex separately used within-source sex ratios to create age-sex-specific estimates. Data covering >25 years were split using an alternative age pattern estimated from external sources.

Clinical informatics data (inpatient admissions, outpatient/GP visits, insurance claims) were also processed. Single-diagnosis inpatient data were adjusted for readmissions, non-primary diagnoses, and outpatient care. Individual-level sources provided fractions of unique cases and ratios (primary/non-primary diagnoses, inpatient/outpatient care). We estimated cause-specific age-sex ratios using MR-BRT for inpatient adjustments. Per capita total inpatient admission rates scaled sources incomplete for the population. Inpatient data were further transformed using a healthcare access and quality index-derived scalar to adjust for location-based access differences. These steps standardize population-level clinical incidence and prevalence estimates⁴¹. For most injuries, prevalence and incidence were modeled using DisMod-MR 2.1 (disease modeling meta-regression), a Bayesian meta-regression tool that generates internally consistent estimates of prevalence, incidence, remission, and mortality by age, sex, location, and year. It estimates these measures for locations lacking raw epidemiological data by cascading estimates through the five-level GBD geographical hierarchy, using data from higher-level locations as priors for lower-level estimation. DisMod-MR 2.1 also leverages location-level covariates to inform estimates in data-sparse areas⁴².

Spatiotemporal Gaussian process regression (ST-GPR)

For select causes, spatiotemporal Gaussian process regression (ST-GPR) replaced DisMod-MR 2.1. ST-GPR applies regression methods to smooth heterogeneous, incomplete data across age, time, and geography. Custom models were developed for causes where neither DisMod-MR 2.1 nor ST-GPR provided adequate estimation of prevalence or incidence³¹. The input data were modeled by using ST-GPR to allow for smoothing over age, time, and location in locations that were missing complete datasets. The approach is a stochastic modeling technique that is designed to detect signals amidst noisy data. It also serves as a powerful tool for interpolating non-linear trends. GPR assumes that the specific trend of interest follows a Gaussian process.

Statistics and reproducibility

Firstly, the single or comprehensive burden of DOH, DOK and DOS (including incidence, prevalence and YLDs) was described, analyzed and visualized according to different SDI levels, gender and age groups.

To determine the direction and magnitude of the time-varying, we calculated the average annual percentage change (AAPC) of the age-standardized incidence rate (ASIR), age-standardized prevalence rate (ASPR), and age-standardized YLDs rate (ASIR) with the corresponding 95% confidence intervals (CI) by the joinpoint model, to quantify the rate trends from 1990 to 2021. A positive AAPC indicates an upward trend, while a negative AAPC indicates a downward trend. We further use maps to visually compare the burden across countries. Subsequently, the age-period-cohort (APC) model was established to explore the effects of age, period, and birth cohort of DOH, DOK, and DOS disease burden. We constructed the APC model by referring to the research methods of previous literature⁴³. To further quantify cross-border inequality, we analyzed the correlation between age-standardized rates and the SDI was calculated. Frontier analysis and calculation of the concentration index were also carried out. Frontier analysis and the concentration index serve as standardized metrics for assessing absolute and relative inequalities, respectively, in the distribution of joint dislocation burden across nations⁴⁴. Specifically, frontier analysis was employed to investigate the association between joint dislocation burden and development status as quantified by SDI. This analytic approach establishes a non-linear epidemiological frontier that identifies the theoretically minimum achievable disease burden for each development level by connecting optimally performing countries that demonstrate the lowest joint dislocation burdens relative to their SDI. The vertical distance from this frontier, termed the “effective difference”, quantifies the unrealized potential for burden reduction—representing the gap between observed burdens and the SDI-adjusted achievable minimum that could potentially be addressed through optimized resource allocation. To operationalize this concept, we implemented non-parametric data envelopment analysis following established methodologies described in prior research^45–47. The gap between a country’s observed YLDs and its frontier is termed the efficiency gap, indicating unrealized health gains relative to its development level. This approach highlights areas for improvement and disparities between countries based on their development status. The concentration index, calculated by integrating the area under the Lorenz curve, compares the cumulative proportion of YLDs with the cumulative population distribution sorted by SDI^48,49. The concentration index ranges from −1 to 1, with negative values indicating the concentration of health outcomes in poorer countries and positive values indicating the concentration of health outcomes in richer countries. A value of 0 indicates exact equality.

To clarify the key factors influencing changes in the disease burden of DOH, DOK, and DOS from 1990 to 2021, a degree of attribution analysis was conducted. This analysis assessed the contribution of each factor⁵⁰. The factors include adverse effects of medical treatment, animal contact, conflict and terrorism, drowning, environmental heat and cold exposure, exposure to forces of nature, exposure to mechanical forces, falls, fire, heat, and hot substances, foreign body, interpersonal violence, other transport injuries, poisonings, police conflict and executions, road injuries, and self-harm. Specific definitions and ICD codes for these factors can be accessed on the website⁵¹.

Finally, we applied a Bayesian APC (BAPC) model with an ensemble nested Pierre-Simon Laplace approximation to forecast the future burden up to 2045⁵². BAPC used integrated nested Laplace approximations (INLA) for full Bayesian inference. BAPC generates age-specific and age-standardized projected rates. This study follows standardized reporting of burden of disease studies (STROBOD) principle, and the STROBE checklist can be found in the attachment⁵³. A p-value of less than 0.05 indicates statistical significance. All statistical analyses and data visualizations were conducted using R software (version 4.3.2).

Ethics

Ethical approval and informed consent were waived as the GBD data is publicly available and does not include identifiable information.

Results

Dislocation of hip

In 2021, significant disparities were observed in ASYR, ASIR, and ASPR for global DOH (Fig. 1A). Incidence, prevalence, and YLDs for DOH increased by 18.37%, 59.20%, and 57.21%, while ASIR, ASPR, and ASYR showed a slight decline (AAPC [95%CI]: −0.71[−0.79, −0.63], −0.52[−0.57, −0.47], and −0.54[−0.59, −0.48]) (Table 1). Males had higher counts for DOH but lower growth rates and AAPC compared to females (Fig. S1A and Supplementary Data 6). DOH had the highest prevalence and YLDs, gradually increasing, with incidence being the lowest (Fig. 2).

Fig. 1. Global age-standardized YLDs rates (ASYR) for joint dislocations, categorized by countries/regions in 2021.

A DOH, B DOK, and C DOS. ASYR age-standardized YLDs rates, DOH dislocation of hip, DOK dislocation of knee, DOS dislocation of shoulder, YLDs years lived with disability.

Table 1.

Global incidence, prevalence, and YLDs of dislocation of hip, knee, and shoulder from 1990 to 2021

	DOH		DOK		DOS
Year	1990	2021	1990	2021	1990	2021
Incidence/1000 [95%UI]	2052.92 [1388.08, 2841.63]	2429.94 [1634.46, 3549.25]	2984.88 [2149.17, 4189.81]	3129.41 [2203.90, 4521.67]	4827.10 [3708.98, 6417.43]	5557.19 [4227.07, 7527.32]
Prevalence/1000 [95%UI]	4603.02 [4169.00, 5299.70]	7328.04 [6470.16, 8688.90]	601.67 [488.68, 760.21]	781.89 [647.14, 955.41]	712.44 [162.24, 1280.87]	822.72 [201.05, 1484.65]
YLDs/1000 [95%UI]	69.43 [35.72, 120.86]	109.15 [57.18, 190.16]	66.62 [42.21, 100.45]	85.53 [55.70, 125.80]	44.18 [8.92, 89.10]	51.02 [10.91, 104.54]
ASIR per 100,000 population [95%UI]	38.29 [26.02, 53.53]	30.63 [20.65, 44.80]	53.40 [38.63, 74.45]	40.24 [28.41, 58.24]	89.85 [68.65, 119.56]	69.96 [53.23, 94.34]
ASPR per 100,000 population [95%UI]	102.42 [93.45, 115.80]	87.30 [76.93, 103.67]	12.02 [9.96, 14.88]	9.63 [7.93, 11.87]	13.27 [3.09, 23.89]	10.35 [2.51, 18.77]
ASYR per 100,000 population [95%UI]	1.53 [0.80, 2.68]	1.30 [0.68, 2.26]	1.32 [0.85, 1.96]	1.05 [0.68, 1.55]	0.82 [0.17, 1.66]	0.64 [0.14, 1.31]
Period	1990–2021
Percentage change in incidence, 1990–2021 [%]	18.37		4.80		15.12
Percentage change in prevalence, 1990–2021 [%]	59.20		29.95		15.48
Percentage change in YLDs, 1990–2021 [%]	57.21		28.38		15.48
AAPC of ASIR [95% CI]	−0.71 [−0.79, −0.63]		−0.93 [−0.99, −0.88]		−0.81 [−1.12, −0.51]
AAPC of ASPR [95% CI]	−0.52 [−0.57, −0.47]		−0.74 [−0.97, −0.52]		−0.85 [−0.89, −0.81]
AAPC of ASYR [95% CI]	−0.54 [−0.59, −0.48]		−0.75 [−0.98, −0.52]		−0.85 [−0.89, −0.81]

AAPC average annual percentage change, ASYR age-standardized YLDs rate, ASPR age-standardized prevalence rate, ASIR age-standardized incidence rate, CI confidence interval, DOH dislocation of hip, DOK dislocation of knee, DOS dislocation of shoulder, UI uncertainty interval, YLDs years lived with disability.

Fig. 2. Global incidence, prevalence, and years lived with disability (YLDs) of joint dislocations, along with the trends in age-standardized rates (ASR), from 1990 to 2021.

A Incidence. B Prevalence. C YLDs. The blue line represents dislocation of hip (DOH), the green line represents dislocation of knee (DOK), and the orange line represents dislocation of shoulder (DOS). ASR age-standardized rates, DOH dislocation of hip, DOK dislocation of knee, DOS dislocation of shoulder, YLDs, years lived with disability.

Males had lower rates across all SDI groups, and the highest YLDs proportion was in the High-middle SDI region, especially for females (Fig. 3). The highest YLDs proportion was in people over 30 years old, peaking in those over 80 (Fig. S5). In most SDI regions, incidence, prevalence, and YLDs for DOH showed an upward trend. Low SDI regions had the most significant increases at 63.01%, 139.12%, and 137.91%, with a rise in ASPR and ASYR (Fig. S2A and Supplementary Data 7). The middle SDI region had the highest counts, and Afghanistan had the highest ASIR(Rate[95%UI] = 157.41[78.59, 322.83]), ASPR(Rate[95%UI] = 612.09[205.96, 1649.36]), and ASYR(Rate[95%UI]  = 9.03[2.61, 24.67]) (Supplementary Data 8).

Fig. 3. Global proportion of years lived with disability (YLDs) for SDI and sex distribution of joint dislocations in 2021.

The blue represents dislocation of hip (DOH), the green represents dislocation of knee (DOK), and the orange represents dislocation of shoulder (DOS). DOH dislocation of hip, DOK dislocation of knee, DOS dislocation of shoulder, YLDs years lived with disability.

APC model forecasts suggest ASIR for DOH will plateau between ages 15 and 24 before declining, while ASPR and ASYR will rise before age 75, peaking in people over 80 (Fig. 4A–C). Males consistently have lower ASR, and ASR is expected to decrease globally (Figs. S6–S8). By 2045, incidence, prevalence, and YLDs are expected to rise to 2,642,471, 10,140,436, and 148,935 cases, respectively, with the largest increase in prevalence among females (Supplementary Data 10).

Fig. 4. Global age-period-cohort model for joint dislocations from 1990 to 2021: age effects.

After considering the confounding effects of period and birth cohort, changes in incidence, prevalence, YLDs across various age groups from 1990 to 2021 and peaks were showed. A–C The changes in the incidence, prevalence, and YLDs of DOH. D–F The changes in the incidence, prevalence, and YLDs of DOK. G–I The changes in the incidence, prevalence, and YLDs of DOS. DOH dislocation of hip, DOK dislocation of knee, DOS dislocation of shoulder, RR risk ratio, YLDs, years lived with disability.

Frontier analysis identified Afghanistan, the Syrian Arab Republic, and Eritrea as having the greatest potential for improvement. Belgium, Finland, and Luxembourg had the poorest DOH control despite high SDI, while Madagascar and Malawi had the best (Fig. 5A and Supplementary Data 9). The concentration index slightly decreased, showing improved equity in 2021 compared to 1990 (Figs. 5B and S9A–C). Falls remained the leading cause of YLDs for DOH in 2021. Unlike DOK, interpersonal violence had a higher attributable fraction, linked to contact sports, highlighting the need for injury prevention policies (Supplementary Data 11).

Fig. 5. Inequality analysis of global joint dislocations based on the social-demographic index (SDI) for incidence, prevalence, and years lived with disability (YLDs), from 1990 to 2021, including frontier analysis and concentration index analysis.

A, B DOH, C, D DOK, and E, F DOS. A, C, E show the frontiers analysis, depicting the relationship between SDI and age-standardized YLDs rates (ASYR), with dots representing individual countries by population size. The red dots show countries with higher rates in 2021 than in 1990, while the blue dots show countries with lower rates in 2021 than in 1990. The black dots mark the top 15 countries with the biggest gap between frontier and the real thing. The red font represents the five countries above the high threshold, where the gap between the frontier and the actual value is the largest, and the blue font represents the five countries below the low threshold, where the gap between the frontier and the actual value is the smallest. Closer to the frontiers (black line), the disease is well controlled; farther away, the less well controlled. B, D, F Show the concentration index analysis. The closer the Hendrik Lorentz line is to the diagonal, the smaller the inequality the farther away from the diagonal, the greater the inequality. The Hendrik Lorentz line below the diagonal shows that the burden is concentrated in rich countries and above it in poor countries. DOH dislocation of hip, DOK dislocation of knee, DOS dislocation of shoulder, SDI social-demographic index, YLDs years lived with disability.

Dislocation of knee

In 2021, significant disparities were observed in the ASYR, ASIR, and ASPR for DOK (Fig. 1B). Global incidence, prevalence, and YLDs for DOK increased by 4.84%, 29.95%, and 28.38%, while ASIR, ASPR, and ASYR declined(AAPC[95% CI]: −0.93[−0.99, −0.88], −0.74[−0.97, −0.52], and −0.75[−0.98, −0.52]) (Table 1). Males had higher DOK counts but lower growth rates and AAPC for ASIR, ASPR, and ASYR compared to females (Fig. S1B and Supplementary Data 6). Incidence and YLDs showed larger changes globally (Fig. 2), with males having higher rates only in the High SDI group (Fig. S3).

Compared to females, males only have higher rates in the High SDI group (Fig. 3). For both genders, the highest proportion was under age 30, peaking under age 5, with males having lower rates across all age groups (Fig. S5). Incidence decreased by 13.95% and 7.98% in high-middle and high SDI regions, respectively, while the burden increased most in low SDI regions. Low-middle and high-middle SDI regions saw annual reductions (Fig. S2B and Supplementary Data 7), but ASIR, ASPR, and ASYR grew in Low SDI regions (Fig. S4 and Supplementary Data 7). Afghanistan had the highest rates for DOK (ASIR: rate[95%UI] = 160.86[79.66, 340.92], ASPR: rate[95%UI] =  41.08[19.92, 87.61], ASYR: rate[95%UI] = 4.35[2.07, 8.73]), while Kiribati had the lowest (ASIR: rate[95%UI] =  15.02[10.64, 20.96], ASPR: rate[95%UI] = 3.94[3.30, 4.77], ASYR: rate[95%UI] = 0.44[0.28, 0.65]) (Supplementary Data 8).

APC model forecasts show ASIR peaking for males at ages 15–19 and females at ages 5–9, then declining, with peaks at ages 75–79 for ASPR and ASYR (Fig. 4D–F). The annual change for DOK is less than 0 for most age groups, indicating a decrease in ASR (Fig. S6a, d, e). After considering the confounding effects of age and birth cohort, people globally during the period 2017–2021 are protected compared to the period 2002–2006 (risk ratio [RR] <1) (Fig. S6b, d, e). Furthermore, after accounting for the confounding effects of age and period, people born during 2017–2021 are protected compared to those born during 1957–1966 (RR < 1) (Fig. S6c, d, e). The global burden is expected to increase from 2022 to 2045, while ASIR, ASPR, and ASYR will decline (Supplementary Data 10).

Frontier analysis identified Afghanistan, New Zealand, and Australia among the top five countries with the most improvement potential, while Belgium and Finland had the greatest gap in DOK control (Fig. 5C and Supplementary Data 9). Madagascar and Malawi had the best control. The concentration index for DOK decreased slightly, indicating reduced inequality, but DOK showed greater disparity favoring wealthier countries (Figs. 5D and S9D–F). Unlike DOH, the top causes of DOK incidence and YLDs in 2021 were falls, road injuries, interpersonal violence, and exposure to mechanical forces. Falls remained the top cause of YLDs. Exposure to mechanical forces, being one of the main causes of DOK, suggests the need for specialized protective policies against external violence injuries (Supplementary Data 11).

Dislocation of shoulder

In 2021, significant disparities in ASYR, ASIR, and ASPR for DOS were observed across 204 countries/regions. The results were shown using color-coded maps (Fig. 1C). From 1990 to 2021, global incidence, prevalence, and YLDs for DOS increased by 15.12%, 15.48%, and 15.48%, respectively, with slight declines in ASIR, ASPR, and ASYR (AAPC[95% CI]: −0.81[−1.12, −0.51], −0.85[−0.89, −0.81], and −0.85[−0.89, −0.81]) (Table 1). Males had higher DOS counts but lower growth rates and AAPC for ASIR, ASPR, and ASYR compared to females (Fig. S1C and Supplementary Data 6).

DOS had the highest incidence and the lowest prevalence and YLDs (Fig. 2), with males having higher rates across all SDI groups (Fig. S3). YLDs for DOS were lowest in all SDI regions, with the highest proportion in females in High-middle SDI regions (Fig. 3). By 2021, YLDs were lowest across all age groups, peaking in the under-5 group. Males had higher rates in all age groups (Fig. S5). Between 1990 and 2021, except for a decrease in the high-middle SDI region, other regions showed increases in incidence, prevalence, and YLDs. The Low SDI region saw the largest increases. AAPC for ASIR, ASPR, and ASYR declined in all SDI regions, indicating a decreasing burden (Figs. S2C; S4 and Supplementary Data 7). ASIR, ASPR, and ASYR in the Low SDI region showed a fluctuating downward trend (Fig. S4 and Supplementary Data 7). New Zealand had the highest ASIR(Rate[95%UI] = 225.18[162.82, 322.78]), ASPR(Rate[95%UI] = 33.01[6.90, 62.39]), and ASYR(Rate[95%UI] = 2.05[0.37, 4.29]), and Kiribati the lowest ASIR(Rate[95%UI] = 23.73[18.34, 31.03]), ASPR(Rate[95%UI] = 3.52[0.86, 6.29]), and ASYR(Rate[95%UI] =  0.22[0.05, 0.44]) (Supplementary Data 8). The age-period-cohort model predicted a rise in ASIR, ASPR, and ASYR from ages 15 to 24, followed by a peak at over 80 years old. Males had lower ASR across all age groups (Fig. 4G–I). Between 1990 and 2021, ASR decreased in most age groups, with increases after age 69 (Figs. S6g–i; S7g–i and S8g–i). By 2045, global DOS cases will increase, while ASIR, ASPR, and ASYR will decrease, especially among males (Supplementary Data 10).

Inequality analysis showed New Zealand, Afghanistan, Australia, Finland, and Andorra had the most potential for improvement. Despite high SDI, countries like Belgium and Switzerland showed poor DOS control. In contrast, low-SDI countries like Madagascar and Malawi had better control (Fig. 5E and Supplementary Data 9). The concentration index slightly decreased in 2021, though DOS showed greater inequality favoring wealthier regions compared to DOH and DOK (Figs. 5F and S9G–I). The study also analyzed 16 injury causes for DOS from 1990 to 2021. Falls, road injuries, and exposure to mechanical forces remained the top causes, with falls leading in YLDs in 2021. Policies targeting these causes remain crucial for reducing the DOS burden (Supplementary Data 11). Unlike DOK, the attributable fraction of exposure to mechanical forces for DOS is higher than that of conflict and terrorism at this time, suggesting that policies addressing mechanical forces can provide effective protection for both DOK and DOS.

Discussions

This study leveraged data from the GBD 2021 study to conduct an extensive and contemporary assessment. It provided a detailed description of the disease burden of DOH, DOK, and DOS, stratified by age, sex, SDI, and geographical region. Additionally, it examined disparities between nations and projected future disease burdens. Unlike previous studies that primarily focused on incidence or prevalence, this analysis utilized YLDs to measure the disease burden, offering a comprehensive perspective for a comprehensive understanding of the impact of joint dislocations, a type of natural injury.

Exploring age and gender differences in patients with DOH, DOK, and DOS helps identify those needing extra care. Since 1990, global ASYR for these dislocations has risen, with similar trends across age, gender, and SDI regions. DOH primarily affects those over 30, while DOK and DOS predominantly affect those under 30, emphasizing the need for targeted interventions: elderly populations for DOH and younger individuals for DOK and DOS. Adult DOH often links to severe hip and femoral neck fractures, while in children, minimal force can cause isolated dislocations due to anatomical differences^54,55. Children’s traumatic hip dislocations are rare, making up less than 5% of cases in specialized centers, but their frequency is significantly higher than in adults^12,56. Similar patterns were observed in this study, with DOH being much more common in those over 80, at seven times the rate of those under 5 years old^12,57,58.

Gender differences in dislocation incidence are notable, with males more prone to dislocations^59–61. The burden of DOK and DOS remains higher in males than females, consistent with prior studies^62–64. For DOH, Brazilian studies show an average age of 34.4 years, with 90.7% being male⁶⁵, while a German study found 79.5% of DOH cases were male, averaging 43 years old⁶⁶. In the U.S., male athletes had higher sports related DOH incidences, with the most cases in adolescents aged 15–19²⁵. Braun et al. noted two peaks in German adolescents for DOH (ages 4–8 and 11–15), with 16% of patients developing avascular necrosis of the femoral head²⁴. This pattern aligns with our findings, showing a rise in DOH between ages 10 and 24, peaking at 79, and rising sharply again in those over 80. DOK mainly affects young adult males, showing increased incidence and prevalence around age 79, peaking in those over 80, with the first YLDs peak at 15–24 years old. The high incidence in young males may be due to their involvement in high-risk jobs and activities, while the elderly, particularly females, also represent a significant risk group due to frailty⁶⁷. For DOS, trends mirror those of DOK. Patrick et al. found the highest incidence in males aged 15–20, with females experiencing a consistent rate throughout life and surpassing males after age 63⁶⁸. Studies by Leroux et al. and Shah et al. corroborate this^69,70. In Iran, DOS is the most common injury, particularly among young males, with rates peaking at 21–30 years before rapidly declining, while female incidence remains low⁶³. The burden of DOS is also higher among U.S. military personnel, with an incidence rate of 169/100,000⁷¹. This may be due to the physically active nature of the population studied by Owens et al., representing individuals engaged in high-risk sports and activities. Elderly females are another major groups at high-risk for DOS, with age-related muscle loss contributing to an increased fall risk^72,73. Gender differences in age-related changes, such as muscle loss, may also play a role^74,75. The study observed that from 1990 to 2021, the burden of joint dislocations in females has increased at a higher rate than in males, potentially due to cultural and social barriers that affect women’s access to healthcare, as suggested by prior research⁷⁶. Increased attention to women’s health and lifestyle factors, such as diet, stress, lack of exercise, smoking, and alcohol consumption, may also influence these trends^77–79. These findings emphasize the importance of tailored interventions and prevention strategies for groups at high-risk for joint dislocation.

The study’s comprehensive analysis of the global burden of joint dislocations reveals the impact of socio-demographic development levels on the uneven distribution of disease burden. The SDI effectively measures socio-demographic development across regions and countries, helping to quantify disparities and understand variations in joint dislocation burden^80,81. High SDI regions typically benefit from longer life expectancies, accelerated aging, and robust healthcare systems, which more accurately capture the true disease burden. These regions, however, face the challenge of an aging population, leading to a heavier healthcare burden for managing conditions like joint dislocations. Conversely, low SDI regions struggle with insufficient healthcare services, limited coverage, and challenges in accessing medical care. This results in potential misdiagnoses and an underestimation of the actual burden of joint dislocations, keeping the disease growth higher for all three dislocation types—DOH, DOK, and DOS. The global distribution of joint dislocations is also influenced by various factors, such as lifestyles, socio-economic conditions, climates, and geographic characteristics. Particularly, regions like Australia, western Asia, the middle East, Central Europe, and Eastern Europe bear a disproportionately high burden, demanding targeted healthcare interventions. In high SDI countries, universal healthcare coverage enables better detection and management of diseases, such as DOS. The relationship between SDI and joint dislocation burden suggests that socio-demographic and economic factors should inform prevention and treatment strategies. Economic status plays a significant role in the incidence of joint dislocations, with both the wealthiest and the poorest populations experiencing high incidence rates⁸². Wealthier individuals can afford timely medical care, while economically disadvantaged populations face a higher risk of traumatic dislocations due to engagement in hazardous work, making them more susceptible to joint injuries. This highlights the need for interventions tailored to both ends of the socio-economic spectrum to address these disparities effectively.

The study’s findings reinforce the critical role that falls play in the occurrence of joint dislocations, such as DOH, DOK, and DOS, particularly as the population ages. Falls are not only the most common cause of these dislocations, but they are also a growing public health challenge as health expectancy continues to increase^83–85. The elderly population is especially vulnerable, and falls represent a key public health concern^86,87. Analyzing our data, we observe significant variations in the impact of these factors across different regions and joint dislocation types. For instance, in high-income regions, falls account for a substantial proportion of joint dislocations, with the rate in DOH/DOK approaching 50k, and an increase of nearly 10k over the past decade. In contrast, DOS shows a lower rate, with growth nearing 10k over the last 30 years. For interpersonal violence, the rate in DOH has risen, whereas it has decreased in DOK and DOS. This trend indicates that interpersonal violence is a more significant factor in hip dislocations than in knee and shoulder dislocations. Research by Zacchilli et al.⁸⁸ and Nordqvist et al.⁸⁹ further underscores the prevalence of falls as a cause of DOS, with indoor falls being particularly common among individuals aged 65 and older. To mitigate the impact of falls, especially in elderly populations, it is essential to implement systemic measures such as regular fall risk assessments and tailored interventions^90–92. Collaboration between communities, healthcare institutions, and care facilities can significantly reduce the incidence of falls and related injuries by introducing preventive strategies like strength training, home modifications, and balance exercises. Interpersonal violence, especially in contact sports, is another major cause of joint dislocations, influenced by biological, behavioral, and regulatory differences between genders⁹³. Basketball, soccer, and contact sports, such as wrestling and hockey, are associated with a significant proportion of sports-related joint dislocations, particularly among males aged 20–24^17,59,93. Moreover, gymnastics, volleyball, and skiing also carry a high risk for joint dislocation, reflecting the broader vulnerability of athletes across various disciplines. In terms of road injuries, traumatic DOH is frequently linked to traffic accidents, particularly those involving high-energy mechanisms that lead to severe injuries. Lima et al.⁶⁵ found that 95% of traumatic DOH cases involved traffic accidents, and studies by Cooper et al.⁹⁴ and Babalola et al.⁹⁵ highlight that even with modern safety devices, road injuries remain a leading cause of severe dislocations, particularly in regions like Nigeria. These findings suggest a pressing need for enhanced road safety measures and policy interventions to better protect individuals from traumatic dislocations in vehicle collisions.

The increasing global burden of joint dislocations, particularly DOH, DOK, and DOS, highlights the critical need for focused rehabilitation strategies to improve patient outcomes and reduce long-term complications. Research by Ma et al. found that, despite emergency interventions, such as closed or open reduction for patients with traumatic hip fractures and dislocations, approximately 24% of cases resulted in poor outcomes, including avascular necrosis, traumatic arthritis, and the need for total hip arthroplasty ⁹⁶. These results emphasize the complexity and challenges of effective rehabilitation for such injuries. Yaari et al. reported favorable outcomes for most patients with anterior hip dislocations following treatment but also noted the occurrence of adverse symptoms, such as heterotopic ossification, recommending timely surgical intervention for DOH when appropriate⁹⁷. Ahmed et al. reinforced the importance of early reduction in traumatic hip dislocations to decrease the incidence of femoral head avascular necrosis, advocating for prompt medical responses to optimize patient prognosis⁹⁸. In terms of DOK, rehabilitation remains a challenging area, particularly after surgery for multiple ligamentous knee injuries, where loss of motion is not uncommon⁹⁹. The literature suggests that early rehabilitation following multiligamentous knee reconstruction might yield slightly better outcomes than delayed rehabilitation, but the improvements are not substantial¹⁰⁰. Furthermore, patellar instability, a frequent cause of knee discomfort, especially in young athletes, has a high recurrence rate ranging from 15% to 71%^29,101–103, underscoring the necessity for targeted preventive strategies and early interventions for groups at high-risk for joint dislocation. The uncertainties surrounding the rehabilitation of knee dislocations, coupled with a lack of consensus on optimal rehabilitation protocols, point to an urgent need for more large-scale, high-quality studies. These studies could help establish evidence-based rehabilitation practices, improve patient outcomes, and reduce disability.

This study has several limitations, primarily inherent to the global burden of disease (GBD) data sources. First, our analysis is heavily dependent on GBD databases, whose accuracy is influenced by the variable accessibility of primary data, such as national registries. This reliance introduces potential biases. For instance, selection bias may arise if data are primarily available from tertiary care centers in urban areas, underrepresenting cases in rural or underserved communities. Additionally, information bias can occur due to inconsistencies in diagnostic practices and coding across different healthcare systems, leading to misclassification or underreporting of joint dislocations. These limitations, including unrecorded cases and a lack of data on specific risk factors, are inherent to the GBD framework, which must contend with the inconsistent availability and variable quality of global epidemiological data^104,105. Known challenges include potential underreporting of non-fatal outcomes like joint dislocations, a lack of data on specific risk factors, and methodological heterogeneity in Source data. To address these challenges, the GBD study employs sophisticated methodological strategies. Disaggregated Bayesian modeling is used to adjust for data gaps and heterogeneity, helping to mitigate the influence of biased source data. Furthermore, the computation of 95% UIs for every estimate quantitatively captures the uncertainty stemming from these biases and data limitations. While these statistical approaches cannot fully substitute for high-quality, standardized primary data, they represent a robust effort to correct for and quantify the impact of potential biases. Ultimately, these data gaps underscore the necessity for continuous strengthening of global injury surveillance systems. In addition, GBD 2021 uses different data sources and different collection and reporting standards, and differences in data collection methods across countries may indeed be a limitation of this study. Differences in healthcare infrastructure, diagnostic practices, and reporting standards across regions can affect data accuracy and completeness. Furthermore, residual instability persists due to incomplete disentanglement of data variation, inevitable exclusion of extant data in each release cycle, and challenges in comprehensively quantifying uncertainty (e.g., 95% UIs) across the estimation pipeline—driving continuous efforts to strengthen data collection systems and analytical methods. From a policy perspective, our findings underscore the urgent need for standardized global registries and enhanced diagnostic protocols for joint dislocations to improve data quality. Public health initiatives should prioritize targeted prevention programs in high-risk regions and populations. Regarding future research, our study confirms that more investigations are needed to refine the injury disease burden estimates. Priority areas include validating our findings with high-quality, primary data from specific regions, conducting studies to better identify and quantify key risk factors, and further refining the disability weights associated with different types and severities of joint dislocations. Therefore, these factors should be taken into account when interpreting study results, as they may lead to data discrepancies. The reliance on models and estimation methods introduces potential biases and uncertainties, and we strive to ensure the authenticity and accuracy of the data analysis process.

Conclusions

Despite the slowed global burden increase of joint dislocations, the overall burden continues to rise, with significant regional disparities. The high burden in low SDI countries calls for urgent targeted interventions. Differences in disease burden by gender and age highlight the need to address the specific needs of various populations. While falls and road injuries are common causes, DOH and DOK are often linked to contact injuries, like sports, and DOS to mechanical force exposure. Strengthening global and regional collaboration, improving access to diagnostic, treatment, and prevention services, and creating targeted interventions for groups at high-risk for joint dislocation are key to reducing the impact of joint dislocations.

Author contributions

Concept and design: JT Huang, HX Tang, ZH Deng, JJ Lin, and XT Zhang. Acquisition, analysis, or interpretation of data: JT Huang, HX Tang, and XT Zhang. Drafting of the manuscript: all authors. Critical revision of the manuscript for important intellectual content: JT Huang, HX Tang, RJ Liang, SC Jia, X Zhang, and JY Su. Statistical analysis: HX Tang, JT Huang. Administrative, technical, or material support: XT Zhang, JJ Lin, ZL Jiao, L Li, and ZH Deng. Supervision: XT Zhang, JJ Lin, ZH Deng, JT Huang, and JY Chen. All authors read and approved the final manuscript.

Peer review

Peer review information

Communications Medicine thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The global burden of disease (GBD) data used in the analyses are available at https://ghdx.healthdata.org/gbd-results-tool. Additional data supporting this article can be obtained upon reasonable request to the corresponding author. And according to the requirements, the author can provide relevant data, and Jianjing Lin will be responsible for providing detailed information in response to this request. The Source data can be viewed through Supplementary Data 1–5. The Source data for Fig. 1 is in Supplementary Data 1. The Source data for Fig. 2 is in Supplementary Data 2. The Source data for Fig. 3 is in Supplementary Data 3. The Source data for Fig. 4 is in Supplementary Data 4. The Source data for Fig. 5 is in Supplementary Data 5.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s43856-026-01418-8.