Trend analysis and Bayesian projection of childhood life-limiting conditions in Korea: insights into incidence and mortality

Abstract

Objective

To estimate long-term trends in the incidence and mortality of life-limiting conditions among children and young people in Korea from 2012 to 2022, and to project these trends through 2035.

Design

A population-based observational study using joinpoint regression for temporal trend analysis and a Bayesian age–period–cohort model with integrated nested Laplace approximation for projections.

Setting

Nationwide data from the Korean National Health Insurance Service and national mortality statistics (2012–2022).

Participants

Children and young people in Korea diagnosed with life-limiting conditions.

Main outcome measures

Age-standardised incidence rates and age-standardised mortality rates per 100 000 population.

Results

Between 2012 and 2022, the age-standardised incidence rate increased from 398.5 to 565.4 per 100 000, while the age-standardised mortality rate decreased from 7.2 to 5.8. In 2022, non-cancer conditions had a 5.43 times higher incidence and 1.16 times higher mortality than cancer-related conditions. Incidence was consistently higher among females. Mortality declined overall; however, the previously declining mortality trend in non-metropolitan areas reversed during the COVID-19 period, although the subsequent increase was not statistically significant. Projections indicate that by 2035, the age-standardised incidence and mortality rates will reach 551.5 and 5.7 per 100 000, respectively.

Conclusions

The increasing incidence and changing mortality patterns of life-limiting conditions among children and young people in Korea indicate growing care needs. Enhancing paediatric palliative care strategies and national surveillance systems is required to address the future burden.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• The number of children and young people living with life-limiting conditions is increasing globally, but most studies focus on prevalence rather than trends in incidence and mortality.

• National, population-based longitudinal evidence remains limited and the burden of non-cancer conditions relative to cancer has not been well characterised.

WHAT THIS STUDY ADDS

• This study provides the first national analysis of long-term trends and Bayesian projections of life-limiting conditions among children and young people in Korea through 2035.

• The data show a marked increase in incidence, a predominance of non-cancer conditions and emerging geographic disparities in mortality during the COVID-19 period.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings support a shift toward integrated paediatric palliative care models and highlight the need for strengthened surveillance systems and regional capacity building to address the projected increase in disease burden.

Introduction

Children and young people (CYP) with life-limiting conditions (LLCs) experience a broad spectrum of diseases, including congenital anomalies, neuromuscular disorders, metabolic diseases and advanced cancers, which ultimately lead to premature death.¹ Advances in medical care and technology have improved survival rates in this population; however, they have also contributed to an increase in the number of children living with chronic and often progressive illnesses over prolonged periods.^{2 3} Due to the young age of patients and the unpredictable course of these diseases, survival outcomes remain difficult to predict.⁴ Even among those who survive, the possibility of complete cure is limited. Given the severity and complexity of their conditions, these children require comprehensive, multidisciplinary management, including early integration of paediatric palliative care (PPC) and ongoing supportive services.^{5 6}

Globally, more than 21 million children are estimated to require PPC each year, with more than 8 million needing some level of specialised care. The demand for PPC varies considerably between countries. For example, estimates range from 120 per 10 000 children in Zimbabwe to 20 per 10 000 in the UK.⁷ Evidence from countries such as England and Australia indicates that the prevalence of LLCs is increasing.^{8 9} Despite this growing burden, epidemiological studies remain limited and have focused predominantly on prevalence rather than incidence or mortality trends.^8–12 Furthermore, the epidemiology of LLCs can differ significantly by geographic region, age group, disease category and socioeconomic status.^8–12

Population-based data are essential for health-system planning, resource allocation, personalised care delivery and policy monitoring. However, to the best of our knowledge, trends in the incidence and mortality of LLCs among CYP worldwide have not been reported. While some studies have addressed prevalence, to date only one has analysed mortality.^8–12 Despite increasing recognition of the need for PPC, population-level monitoring of LLCs remains insufficient in many countries. In particular, the long-term trends in incidence and mortality that are necessary to evaluate the impact of health policies and anticipate future service needs are rarely documented. Without timely and accurate national estimates, it is difficult to assess whether medical advances and health policies have improved outcomes for CYP with LLCs. This underscores the need for continuous, nationally representative research that provides annual estimates of LLC incidence and mortality.

In South Korea, the healthcare system provides universal coverage through the National Health Insurance Service (NHIS), which enables the collection of near-complete population-level data. Over the past decade, policy efforts have expanded coverage for rare and severe paediatric conditions, including high-cost diagnostics, orphan drugs and specialised services. However, challenges remain in delivering coordinated, multidisciplinary care for children with LLCs, particularly in the context of increasing demand for PPC and long-term support. A clearer understanding of disease burden is necessary to guide effective health policy and clinical service delivery.

This study estimates the national incidence and mortality trends of LLCs among CYP in Korea from 2012 to 2022, and projects these trends through 2035. The findings will provide evidence to support paediatric health policy formulation, improve resource planning and inform the development of early palliative-care services.

Methods

Study population and case definition

Incidence of cases

Using claims data from the NHIS, Korea’s single-payer health insurance system. We constructed a nationwide cohort of CYP aged ≤24 years who were newly diagnosed with LLCs between 2012 and 2022. To ensure incident case ascertainment, individuals with any claims for LLCs prior to 2011 were excluded. Both inpatient and outpatient records from general healthcare institutions were included. For individuals with multiple claims in a given year, only the earliest diagnosis date was used to define the incident case (online supplemental figure 1). Eligible individuals were selected by excluding those whose annual qualification records could not be matched across the study period.

Deceased cases

Mortality data were obtained from the microdata of Statistics Korea’s Cause of Death Statistics,¹³ which are primarily based on death certificates, supplemented with information from various administrative datasets, including health insurance and cancer registry records. CYP aged ≤24 years whose cause of death was attributed to a LLC between 2012 and 2022 were included.

Main outcomes

Age-standardised incidence rates and age-standardised mortality rates

We examined the incidence and mortality of CYP with LLCs in Korea between 2012 and 2022. LLCs were defined using the directory of LLCs based on International Classification of Diseases, 10th Revision (ICD-10) codes. A previously adapted framework mapping ICD-10 codes to the sixth Korean Standard Classification of Diseases codes for Korean National Health Insurance claims data was used in this study (see online supplemental table 1).^{10 14}

Outcomes were stratified by sex, age group (0–4, 5–9, 10–14, 15–19 and 20–24 years) and cancer presence (cancer vs non-cancer) to examine subgroup-specific patterns. Disease categories were grouped into cancer-related and non-cancer-related LLCs.

Statistics analysis

Crude and age-standardised rates

Crude incidence rates and crude mortality rates were calculated by dividing the number of LLC cases by the mid-year population of CYP in each calendar year. Mid-year population estimates were obtained from the Korean Statistical Information Service and stratified by sex and 5-year age groups. To enable temporal and international comparisons, age-standardised incidence rates (ASIRs) and mortality rates (ASMRs) were computed using the WHO 2000–2025 standard population for five age groups (0–4, 5–9, 10–14, 15–19 and 20–24).¹⁵ All rates presented are per 100 000 population.

Trends analysis: joinpoint regression

Temporal trends in LLC incidence and mortality from 2012 to 2022 were evaluated using joinpoint regression analysis. Annual percentage changes (APCs) and average APCs (AAPCs), along with 95% CIs, were estimated to assess significant shifts in ASIRs and ASMRs. The analysis was conducted using the Joinpoint Regression Program (V.5.3.0, November 2024; National Cancer Institute (NCI), USA).¹⁶ A maximum of one joinpoint was specified, allowing for up to two distinct trend segments, with model selection based on permutation tests.

Bayesian projection modelling

The Bayesian age-period-cohort (BAPC) model was implemented to project future incidence and mortality rates of childhood LLCs in Korea. This modelling approach quantifies long-term trends by decomposing log-transformed rates into additive age, period and cohort effects, while accounting for uncertainty in each component. The model is expressed as:

$${\displaystyle {\displaystyle \begin{array}{l}\log \left({\lambda }_{i,j}\right)=\mu +{\alpha }_i+{\beta }_j+{\gamma }_k\end{array}}}$$

where λ represents the incidence or mortality rate, μ is the overall intercept, and α_i, β_j, and γ_k represent the effects of age, period and cohort, respectively. Alternative random-walk specifications (RW1 and RW2) were evaluated for age, period and cohort effects. The final model was selected based on the deviance information criterion and the RW1 specification was adopted for the final projection analyses. The integrated nested Laplace approximation was used to estimate marginal posterior distributions of model parameters and generate posterior predictive distributions for future incidence and mortality. Projected values were reported with 95% credible intervals (CrIs) to reflect the uncertainty of long-term projection. Rates for 2023 through 2035 were estimated using historical data from 2012 to 2022. The BAPC model was implemented using the R package BAPC. Projected values are presented as incidence and mortality rates, along with corresponding 95% CrIs.^17–19

All statistical analyses were performed using R statistical software (V.4.4.2; R Core Team, Vienna, Austria), Joinpoint Regression Program Software (V.5.3.0) and SAS V.9.4 (SAS Institute). A two-tailed significance level of α=0.05 was applied for all tests.

Patient and public involvement

Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this research. This study was based on secondary analysis of anonymised administrative data, and no direct patient or public involvement was feasible or appropriate.

Results

ASIR and ASMR

In 2022, 59,621 CYP in Korea were newly diagnosed with LLCs (table 1), with an ASIR of 565.4 per 100 000 population (online supplemental figure 2). This represents an increase from 2012, when 55 895 cases were reported and the ASIR was 398.5. Among LLCs, the ASIR for cancer was 74.6, while that for non-cancer conditions was substantially higher at 490.8. In the same year, 896 deaths occurred among CYP with LLCs (table 1), with an ASMR of 5.8 per 100 000 population (online supplemental figure 3). This was a decrease from 2012, which recorded 975 deaths and an ASMR of 7.2. The ASMR for cancer was 2.3 and that for non-cancer conditions was 3.5. The ASIR increased sharply during the COVID-19 pandemic (2020–2021), particularly among those with non-cancer conditions, accompanied by an increase in ASMR. ASIR and ASMR were similar across sex and metropolitan area of residence.

Table 1. Trends of incidence and mortality per 100 000 children and young people with life-limiting conditions in Korea.

	Incidence							Mortality
	Case			CR			Relative difference	Case			CR			Relative difference
	2012	2017	2022	2012 (a)	2017	2022 (b)	(b–a)/a*100	2012	2017	2022	2012 (a)	2017	2022 (b)	(b–a)/a*100
Total	55 895	56 019	59 621	386.3	418.8	524.5	35.8	975	755	586	6.7	5.6	5.2	−22.4
Age
0–4	17 144	18 950	16 830	737.6	885.0	1138.1	54.3	383	292	199	16.5	13.6	13.7	−17.0
5–9	5637	5184	6437	240.3	220.0	297.2	23.7	97	84	56	4.1	3.6	3.4	−17.1
10–14	7252	5148	7153	238.6	219.4	303.4	27.2	115	80	86	3.8	3.4	3.2	−15.8
15–19	12 151	10 694	11 189	346.3	352.5	478.0	38.0	184	128	87	5.2	4.2	3.7	−28.8
20–24	13 711	16 043	18 012	421.5	458.5	595.7	41.3	196	171	151	6.0	4.9	4.8	−20.0
Sex
Male	30 333	29 545	30 023	400.6	424.5	511.8	27.8	581	447	335	7.7	6.4	5.7	−26.0
Female	25 562	26 474	29 598	370.6	412.5	538.1	45.2	394	308	251	5.7	4.8	4.6	−19.3
Type of disease
Cancer	7953	8036	9278	55.0	60.1	81.6	48.4	465	351	271	3.2	2.6	2.4	−25.0
Non-cancer	47 942	47 983	50 343	331.3	358.7	442.9	33.7	510	404	315	3.5	3.0	2.8	−20.0
Residence
Metropolitan	24 454	23 750	25 843	377.2	406.8	535.1	41.9	440	314	245	6.8	5.4	5.1	−25.0
Non-metropolitan	31 265	32 242	33 776	391.5	427.7	516.6	32.0	535	441	341	6.7	5.8	5.2	−22.4
Missing	176	27	2	–	–	–	–	–	–	–	–	–	–	–

Residence was classified as either metropolitan or non-metropolitan, with metropolitan including special and general metropolitan city regions (Seoul, Busan, Daegu, Incheon, Gwangju, Daejeon, Ulsan and Sejong).

CR, crude incidence rate.

Joinpoint analysis

From 2012 to 2022, the ASIR of CYP with LLCs had an overall upward trend, while the ASMR of CYP declined (table 2). The AAPC was 3.95% (95% CI 2.86% to 5.77%) for ASIR and –2.74% (95% CI –3.72% to –1.74%) for ASMR. Females and metropolitan residents experienced steeper increases in ASIR and more rapid declines in ASMR. For females, the ASIR AAPC was 4.83% (95% CI 3.53% to 7.04%) and the ASMR AAPC was –3.21% (95% CI –5.11% to –1.24%). For metropolitan residents, the ASIR AAPC was 4.35% (95% CI 2.69% to 6.81%) and the ASMR AAPC was –3.40% (95% CI –5.45% to –1.35%).

Table 2. AAPCs in ASIR and ASMR in 2012–2022.

	ASIR		ASMR
	AAPC (95% CI)	P value	AAPC (95% CI)	P value
Total	3.95 (2.86 to 5.77)	<0.001	−2.74 (−3.72 to −1.74)	<0.001
Sex
Male	3.15 (1.89 to 5.27)	<0.001	−2.36 (−3.25 to −1.45)	<0.001
Female	4.83 (3.53 to 7.04)	<0.001	−3.21 (−5.11 to −1.24)	<0.001
Type of disease
Cancer	4.07 (3.12 to 5.64)	<0.001	−3.08 (−4.95 to −1.15)	<0.001
Non-cancer	3.93 (2.63 to 6.06)	<0.001	−2.49 (−3.75 to −1.21)	<0.001
Residence
Metropolitan	4.35 (2.69 to 6.81)	<0.001	−3.40 (−5.45 to −1.35)	<0.001
Non-metropolitan	3.68 (2.52 to 5.68)	<0.001	−1.83 (−2.69 to −1.28)	<0.001

Residence was classified as either metropolitan or non-metropolitan, with metropolitan including special and general metropolitan city regions (Seoul, Busan, Daegu, Incheon, Gwangju, Daejeon, Ulsan and Sejong).

AAPC, average annual percentage change; ASIR, age-standardised incidence rate; ASMR, age-standardised mortality rate.

A change point was observed around 2014 across all ASIR subgroups, with a declining trend from 2012 to 2014 followed by an increase thereafter (online supplemental figure 4). These patterns were consistent regardless of sex, cancer status or metropolitan residency. ASMR declined steadily from 2012 to 2022 (online supplemental figure 5), with downward trends observed across both sexes and disease types. The only exception was in non-metropolitan areas, where the previously declining ASMR trend reversed after 2020, with ASMR increasing during the COVID-19 period (APC, 3.21%; 95% CI −1.71% to 6.41%), although this increase was not statistically significant.

Future projections

We projected ASIR and ASMR trends through 2035 using the BAPC model. The ASIR is projected to decline slightly and then remain relatively stable, reaching approximately 551.5 (95% CrI 193.6 to 909.4) per 100 000 population by 2035. Stratified by sex, the projected ASIR is 551.6 (95% CrI 187.5 to 915.7) for males and 560.4 (95% CrI 187.2 to 933.7) for females (figure 1). In contrast, the ASMR is projected to remain relatively stable, reaching 5.7 (95% CrI 3.9 to 7.5) per 100 000 population by 2035, with sex-specific estimates of 6.4 (95% CrI 4.7 to 8.1) for males and 5.0 (95% CrI 3.2 to 6.9) for females (figure 2).

Figure 1. Projection of age-standardised incidence rates in children and young adults with life-limiting conditions in Korea by sex. Panels show the results for (A) the overall population, (B) males and (C) females. Population projections for 2023–2035 are based on the 2022 Statistics Korea Projections (2024), using the medium variant scenario as of 1 July each year.

Figure 2. Projection of age-standardised mortality rates in children and young adults with life-limiting conditions in Korea by sex. Panels show the results for (A) the overall population, (B) males and (C) females. Population projections for 2023–2035 are based on the 2022 Statistics Korea Projections (2024), using the medium variant scenario as of 1 July each year.

By cancer presence, the ASIR is projected to reach 72.9 (95% CrI 22.2 to 123.7) for cancer and 475.0 (95% CrI 179.3 to 770.6) for non-cancer LLCs (online supplemental figure 6). The ASMR is expected to remain relatively stable in both groups, reaching 2.3 (95% CrI 1.5 to 3.1) for cancer and 3.5 (95% CrI 2.6 to 4.4) for non-cancer LLCs per 100 000 population by 2035.

Discussion

The incidence of LLCs among CYP in Korea increased from 2012 to 2022, while mortality gradually declined. In 2022, the incidence of non-cancer LLCs in CYP was 5.43 times higher, with 1.16 times more deaths, than that of cancer-related conditions. The incidence among females increased over time. In contrast, the previously declining mortality trend among CYP residing in non-metropolitan areas was not sustained during the COVID-19 period. Based on projections using historical data, the incidence and mortality rates of LLCs in CYP are expected to reach 551.5 and 5.7 per 100 000 population, respectively, by 2035.

Understanding trends in both incidence and mortality is essential for designing effective policies for CYP living with LLCs. While prevalence reflects cumulative burden, incidence and mortality offer dynamic insights. From a policy and planning perspective, incidence is critical for forecasting future resource needs, as it captures the inflow of new cases into the healthcare system and is sensitive to changes in diagnostic practices and service access. Mortality reflects disparities in outcomes and the performance of the health system. It also reflects disease severity, care quality and access to interventions and may increase with expansion of PPC programmes. Neither measure alone provides a complete picture. A comprehensive interpretation that considers both—alongside the evolving healthcare landscape—is necessary. By focusing on both indicators, this study provides a more nuanced understanding of the burden of LLCs in Korea, supporting evidence-informed strategies.

LLCs should be recognised as a critical component of public health in CYP. In Korea, cancer is the leading cause of death among children aged 1–9 years. Among those aged 10–29 years, cancer ranks second and cardiovascular disease ranks fourth.²⁰ Given the ongoing decline in the paediatric population, the relative burden of cancer and LLCs more broadly is becoming increasingly significant. This underscores the need for both curative and palliative approaches that address the complex health needs of CYP living with LLCs. To the best of our knowledge, this is the first study to examine national trends in the incidence and mortality of LLCs among CYP. Although previous studies have focused on prevalence, the trends observed here align with previous findings. A supplementary cross-sectional analysis (online supplemental table 2) confirmed this study’s consistency with previous reports of increasing prevalence. For example, studies in the UK and Queensland, Australia, reported increases in prevalence per 10 000 population of 25–32 during 2000/2001–2009/2010 and of 35.2–43.2 in 2016, respectively.^{8 9 12} A previous study in Korea reported an age-standardised prevalence of 955.1 per 100 000 in 2016,²⁰ and our analysis demonstrated a continued upward trend, reaching 1088.1 in 2020.

There are notable differences in the key epidemiological characteristics of LLCs across countries. In England, disparities have been observed based on ethnicity and socioeconomic status.⁹ In Queensland, increases were most prominent among infants and indigenous children.⁸ In a study in the Basque region of Spain, prevalence was higher among males, infants under 1 year of age and socioeconomically disadvantaged groups.¹¹ In a Korean study, cancer was the most common LLC diagnosis, followed by neurological, cardiovascular and renal diseases; the highest burden was observed among infants under 1 year of age.¹⁰ These differences highlight the importance of country-specific epidemiological assessments to inform tailored healthcare strategies for CYP with LLCs.

Trend analysis revealed a significant inflection point in 2014, marking the onset of an increasing trend in the incidence of LLCs among CYP. This upward shift accelerated around 2016, coinciding with major health-policy reforms in Korea that aimed to expand insurance coverage and improve access to diagnostic services. In 2013, the government introduced a policy to enhance financial protection for four major disease groups (cancer, cardiovascular disease, cerebrovascular disease and rare incurable diseases) by classifying them as essential benefit items and reducing patient copayment rates to 5%–10%. Coverage began with ultrasound and cardiac-related MRI examinations in 2013, expanded in 2014 to include high-cost anticancer drugs and advanced diagnostic procedures and further extended in 2015 to radiotherapy and genetic testing. The Rare Disease Management Act, enacted in December 2015, established a legal foundation for rare disease policy and enabled partial insurance coverage for orphan drugs. From 2017 onward, the policy framework expanded beyond the initial disease categories to strengthen overall benefit coverage.²¹ These policy changes likely expanded access to diagnostic evaluation and treatment for serious paediatric conditions and may have contributed to improved case ascertainment. Although a true increase in disease occurrence cannot be excluded, expanded insurance coverage for diagnostic testing may partly explain the observed increase in incidence.

The COVID-19 pandemic likely disrupted healthcare access for CYP with LLCs, particularly in non-metropolitan areas with limited resources. Previous studies have shown that the pandemic exacerbated pre-existing regional socioeconomic inequalities in mortality, with disadvantaged and rural areas experiencing greater increases in mortality than metropolitan regions.²² Such disparities may have amplified the vulnerability of CYP with LLCs, who require continuous access to specialised multidisciplinary care. Restricted utilisation in 2020 may explain the temporary decrease in incidence, followed by a sharp rebound in 2021, possibly reflecting delayed diagnoses and a backlog of cases.^23–25

Additionally, an unexpected increase in mortality among CYP with LLCs was observed in 2018, particularly among those with non-cancer conditions. Although the underlying causes remain unclear, this inflection may reflect the influence of concurrent changes in policy and the healthcare system. The enforcement of Korea’s Act on Hospice and Palliative Care and Decisions on Life-Sustaining Treatment for Patients at the End of Life in August 2017, along with the launch of the PPC pilot programme in 2018, may have led to changes in clinical decision-making, particularly regarding aggressive interventions. These transitions, along with efforts to expand PPC infrastructure and services, could have contributed to shifts in mortality patterns. This observation underscores how changes in policy environments can directly or indirectly influence epidemiological trends, highlighting the need for careful interpretation and adaptive policy planning. These healthcare system changes, together with the ongoing expansion of PPC infrastructure and services, may also have altered patterns of end-of-life care and mortality reporting.²⁶ This observation underscores how changes in healthcare policy and service delivery may directly or indirectly influence epidemiological trends.

Sex differences in disease distribution may partly explain the greater increase in incidence and larger decline in mortality observed among females. In our study, incidence increased among females aged 20–24 years, largely driven by cancer, while respiratory, gastrointestinal and metabolic diseases also increased during the COVID-19 period. The increase in cancer incidence among young females may be understood in the context of known sex differences in cancer distribution. Thyroid, breast, cervical, ovarian and corpus uteri cancers are among the most common cancers in females aged 15–34 years, with some cancer types showing a higher incidence among females than males, suggesting differences in both cancer patterns and incidence burden between sexes in this age group.²⁷ In addition, females have consistently been reported to have a higher risk of post-COVID-19 conditions than males, which may provide context for the increased incidence of several conditions observed among females during the COVID-19 period in our study.^{28 29} In contrast, the larger decline in mortality among females may reflect differences in the distribution of diseases with a high mortality burden. We found higher mortality among males for cancer and several neurologic, neuromuscular and cardiovascular conditions. Previous studies have reported sex-related differences in childhood cancer survival outcomes, suggesting that males may have poorer survival outcomes for several childhood and adolescent cancers.^{30 31} Furthermore, certain neurologic and neuromuscular disorders included in our study, including X-linked disorders, are more commonly observed in males due to sex-linked inheritance patterns and may have contributed to the observed sex differences in mortality trends.^{32 33} However, the mechanisms underlying these sex-specific patterns remain unclear and warrant further investigation.

CYP with LLCs continue to experience disproportionately high mortality rates. A UK-based study reported that children with LLCs admitted to paediatric intensive care units had a higher in-hospital mortality rate (5.4%) compared with those without LLCs (2.7%) and were 2.5 times more likely to die within 1 year after discharge.³⁴ In a study in the Basque region of Spain, the mortality rate among children with LLCs was reported to be 0.6 per 10 000 population.¹¹ In Korea, although the overall mortality rate among CYP has steadily declined, LLCs remain a leading cause of death, accounting for a substantial proportion of paediatric mortality. However, rather than a numerical interpretation, this requires a nuanced understanding that incorporates both clinical complexity and the evolving policy environment. In particular, unlike in adults, decisions regarding limitation or withdrawal of treatment in children involve more complex ethical and social considerations. For example, life-prolonging treatments and palliative care may be delivered concurrently. CYP receiving PPC often experience higher mortality rates, which may reflect deliberate clinical choices rather than treatment failure. Rather than focusing solely on reducing mortality as a policy goal, it is essential to interpret mortality patterns in light of differences in treatment approaches. Accordingly, tailored care planning that reflects disease characteristics, family values and clinical context is essential. Such an integrated approach supports health policies that move beyond mortality as a singular outcome.

Projection analyses imply that if the current trends of increasing incidence and declining mortality persist, the overall prevalence of LLCs in CYP will continue to grow. These findings are consistent with the only study to date that has projected future LLC prevalence. In England, the projected prevalence by 2030 is estimated to be 67.0–84.2 per 10 000 population.¹² In a related context, this trend implies an expanding population of CYP living longer with complex chronic conditions (CCCs), particularly non-cancer diagnoses. A study in Spain reported that 72% of CYP with LLCs also had CCCs.¹¹ In Korea, previous research found that although overall paediatric mortality declined between 2005 and 2014, mortality related to CCCs remained substantial; the prevalence and survival duration of children with non-cancer CCCs increased.³⁵

As the population of CYP living with LLCs continues to grow, demand for comprehensive and integrated PPC tailored to their needs will increase. Such services should not focus solely on patients with cancer but must also reflect the increasing numbers of those with non-cancer conditions. Developing and implementing multidisciplinary palliative care models is essential not only to improve quality of life but also to reduce unnecessary hospitalisations and healthcare costs.³⁶ Prioritising the development and evaluation of tailored palliative care frameworks is vital to meet the growing and complex needs of CYP with LLCs and to address current gaps in service provision.

Several countries have established PPC networks to support standardised practice, interdisciplinary collaboration and workforce development through structured national or regional systems. The UK operates regional children’s palliative care networks, coordinated by Together for Short Lives, to promote integrated services and consistent clinical standards.^{37 38} Australia adopts a hub-and-spoke model under its National Palliative Care Strategy to link tertiary centres with local providers.³⁹ The United States established the Palliative Care Quality Network, a hospital-led collaborative for data sharing, real-time benchmarking and quality improvement.⁴⁰ In Korea, a PPC programme has been implemented through a government-led pilot programme since 2018, providing services by multidisciplinary teams. Integration with clinical-data networks may be needed to facilitate evidence-based quality improvement, enhance continuity of care across settings and better address the complex needs of CYP with LLCs.

This study had several strengths. We used comprehensive National Health Insurance claims data and death records covering virtually all CYP in Korea, allowing robust population-level analysis of incidence and mortality trends over a 10-year period. The incorporation of Bayesian projection modelling provided forward-looking estimates that are valuable for healthcare system planning. However, limitations include reliance on administrative claims data based on diagnostic codes, which may result in misclassification or underreporting of rare or complex conditions and introduce potential inaccuracies in disease classification. The absence of detailed clinical information limits the ability to assess disease severity or functional status, and socioeconomic information (eg, income, education or deprivation indices) was not available in this study. Additionally, healthcare disruptions during the COVID-19 pandemic may have influenced incidence and mortality trends, particularly in 2020–2021. Most international comparisons presented in the Discussion are based on prevalence estimates, whereas this study is focused on incidence and mortality; therefore, direct comparisons may be limited.

Our findings highlight the urgent need for policy action. First, increasing incidence and persistent mortality, especially in non-cancer groups, underscore the importance of early detection and systematic surveillance. Second, expanding access to specialised PPC and embedding multidisciplinary models into routine paediatric care can help meet the growing demand. Third, sex-specific and diagnosis-specific disparities point to the need for tailored care pathways. Finally, sustained investment in paediatric healthcare infrastructure, workforce training and national data systems is critical to support long-term planning and ensure access across regions and populations.

Conclusions

This study provides the first comprehensive, population-based assessment of long-term trends in the incidence and mortality of LLCs among CYP in Korea. The findings reveal a sustained increase in the incidence of LLCs since 2014, with a particularly notable increase in non-cancer conditions. Although overall mortality has decreased, an increase was observed in 2018, especially among females and those with non-cancer LLCs. These patterns may reflect changes in diagnostic practices, healthcare accessibility and disease epidemiology; however, further investigation is needed to clarify the contributing factors.

By incorporating Bayesian projections, this study offers critical insights into the future burden of LLCs. Policy efforts should prioritise the expansion of palliative and multidisciplinary care and ensure equitable access to services. Enhancing national data systems and developing longitudinal surveillance will be essential to support evidence-based planning and improve outcomes for this vulnerable population.

Data availability statement

Data may be obtained from a third party and are not publicly available.