Introduction
Liver disease and diabetes mellitus (DM) are two major public health concerns that increasingly overlap, worsening both morbidity and mortality in the United States. Liver diseases range from nonalcoholic fatty liver disease (NAFLD) and viral hepatitis to cirrhosis and hepatocellular carcinoma (HCC), many of which are strongly influenced by metabolic dysfunction [1, 2]. Meanwhile, diabetes especially type 2 continues to affect millions of Americans and plays a major role in both microvascular and macrovascular complications [1, 2]. When these two conditions co-occur, they can accelerate liver damage, raise the risk of decompensation, and significantly reduce survival outcomes [3].
The connection between liver disease and DM is well established, with insulin resistance, chronic inflammation, and lipid disturbances contributing to disease progression [4, 5]. Diabetes often exacerbates liver conditions, raising the risk of cirrhosis, HCC, and liver-related mortality, particularly in NAFLD, which is now the most common cause of chronic liver disease in Western countries [4–6]. In fact, type 2 diabetes has been found to be an independent predictor of both cirrhosis and HCC in patients with either alcoholic liver disease or NAFLD [6]. Socioeconomic inequalities, lifestyle factors such as poor diet or alcohol use, and limited access to healthcare services further worsen outcomes, particularly among minority populations and rural communities [7, 8].
Despite increasing awareness, there is still a lack of comprehensive national data examining trends in deaths where All-LD and DM were listed as contributing causes over time. Many earlier studies have treated these conditions in isolation or have not analyzed mortality differences across demographic subgroups. With the added disruptions of the COVID-19 pandemic marked by deferred routine care and disproportionate impacts on vulnerable populations the need to study mortality trends and disparities is even more urgent [9, 10].
Therefore, this study aims to analyze national trends in deaths where liver disease and diabetes were listed as contributing causes among U.S. older adults aged 65 and older from 1999 to 2020. We specifically examined death certificates where these conditions were listed as contributing causes to capture the synergistic mortality burden of these comorbidities.
Methods
Study design
The data utilized for this study were sourced from the CDC-WONDER database, a comprehensive repository of death certificates. The dataset spanned from 1999 to 2020. It was used to examine deaths with All-LD and Diabetes Mellitus (DM) listed as causes, employing precise International Classification of Diseases, Tenth Revision (ICD-10) codes [11]. This extensive compilation encompasses death certificates from all 50 states and the District of Columbia. It is an invaluable resource for analyzing mortality trends on various diseases among adults aged 65 and older.
Furthermore, this study utilized de-identified datasets accessible to the public and generated by governmental organizations. Given the characteristics of this data, obtaining institutional review board (IRB) approval was not mandatory. Nevertheless, the research diligently complied with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [12], guaranteeing a high scientific integrity and reliability standard in our findings.
Study cohort
The study encompassed older adults aged 65 years and above. We identified records from the Multiple Cause of Death database where both DM (ICD-10 codes E10–E14) and All-LD (ICD-10 codes K70–K76) were reported on the death certificate. We included cases where these conditions were listed as either the underlying cause or a contributing factor. Consequently, this analysis captures the intersection of both comorbidities regardless of which was coded as the primary underlying cause. This comprehensive approach was chosen because these chronic conditions often act synergistically to accelerate mortality, even when not listed as the immediate underlying cause. Cases in which deaths were not reported to state registries were excluded from the analysis.
To evaluate whether the observed trends in co-occurring mortality were driven by broad changes in single-disease mortality, we performed a supplementary analysis of Diabetes Mellitus and All Liver Disease independently. We queried the CDC-WONDER Multiple Cause of Death database to extract two additional datasets for U.S. older adults (aged ≥ 65) from 1999 to 2020: one identifying all deaths with Diabetes Mellitus (ICD-10 codes E10–E14) listed as a cause, and another identifying all deaths with All Liver Diseases (ICD-10 codes K70–K76) listed as a cause, regardless of comorbidities. Age-adjusted mortality rates (AAMRs) were calculated for each condition separately to allow for a descriptive comparison of their independent trajectories against the synergistic mortality trends of the co-occurring cohort.
Data extraction
This study extracted a dataset that comprises essential mortality factors, including population size, year of death, geographic location, demographic characteristics, urban-rural classification, regional distribution, and state categories. The demographic data encompasses key variables such as age and race/ethnicity. Sex was also extracted for stratification. Death locations are classified into various healthcare settings, including outpatient facilities, emergency rooms, and inpatient hospitals, while differentiating between deaths that occurred upon arrival or had an unknown status, as well as those that transpired at home, in hospice, nursing homes, or long-term care facilities. Race and ethnicity classifications include Hispanic, White, Black, Asian or Pacific Islander and American Indian. These parameters have been validated through an analysis of the CDC WONDER database, which enables a nuanced understanding of mortality trends. To ensure precise assessment of the population, we employed the National Center for Health Statistics Urban-Rural Classification Scheme, which classifies counties as urban (which includes large central metropolitan, large fringe metropolitan, medium metropolitan, and small metropolitan) or nonmetropolitan (encompassing micropolitan and noncore areas) based on the 2013 U.S. Census. Furthermore, utilizing definitions from the U.S. Census Bureau, regions were categorized into the Northeast, Midwest, South, and West, thereby enhancing the specificity and relevance of our findings [13, 14]. These robust classifications furnish critical insights into mortality patterns across diverse populations.
Statistical analysis
To examine nationwide mortality trends, the study comprehensively analyzed the crude mortality rate and the age-adjusted mortality rate (AAMR) per 100,000 individuals. This process involved meticulous calculations of the total fatalities associated with DM within the study population for each year considered. Adhering to established research methodologies, the AAMR was computed by standardizing these DM-related deaths against the 2000 U.S. population while generating 95% confidence intervals to ensure precision. To determine the annual percent change (APC) along with its corresponding 95% confidence interval for AAMR, the JoinPoint Regression Program (Joinpoint V 4.9.0.0, National Cancer Institute, Bethesda, MD, USA) was utilized [15]. Applying AAMRs facilitated meaningful mortality rate comparisons across diverse populations and historical periods. The study elucidated significant mortality trends by analyzing these rates and underscored notable fluctuations over time using log-linear regression models. This research employs AAMRs, APCs, and AAPCs to identify patterns and disparities in mortality across various demographics as done in previous studies [16, 17].
Results
Diabetes mellitus and all liver disease-related yearly pattern for AAMRs
From 1999 to 2020 there were 95,870 All-LD + DM–related deaths in U.S. adults, corresponding to an overall AAMR of 10.2 per 100,000 (95% CI, 9.91–10.53). Deaths rose from 3,227 in 1999 to 8,459 in 2020. The overall AAMR was relatively flat through 2014 (9.3 in 1999 to 10.1 in 2014), increased to 12.6 in 2018, and surged to 15.3 in 2019; joinpoint APCs were 0.08% per year from 1999 to 2014, 5.47% from 2014 to 2018, and 12.42% from 2018 to 2020 (all p < 0.001 for the post-2014 segments). Place of death was most often medical facilities (41,115; 41.85%), followed by home (29,134; 29.65%), nursing homes/long-term care (17,208; 17.51%), hospice facilities (4,957; 5.05%), and other locations (5,836; 5.94%). Home deaths increased markedly over time (724 in 1999 to 3,321 in 2020), while medical-facility deaths also grew (1,741 to 3,127); hospice deaths were rare early on and reached 604 by 2020 (Table 1; Fig. 1, Supplementary Fig. 1, Supplementary Tables 1, 2 and 3).
Table 1.
Demographic characteristics of all liver disease and diabetes mellitus-related mortalities among older adults in the USA from 1999 to 2020
Variable
All-LD and Diabetes Mellitus-related Deaths
AAMRs / CMRs (95% CI) per 100,000
Overall Population
95,870 (100%)
10.2 (9.91–10.53)
Sex
a
Male
50,881 (53.07%)
12.4 (11.91–12.95)
Female
44,989 (46.93%)
8.51 (8.15–8.9)
Census Region
a
Northeast
15,239 (15.9%)
8.55 (7.90–9.19)
Midwest
20,141 (21.01%)
9.70 (9.04–10.31)
South
35,925 (37.47%)
10.22 (9.71–10.73)
West
24,565 (25.62%)
12.23 (11.48–12.96)
Race / Ethnicity
a
NH American Indian or Alaska Native
1350 (1.66%)
27.36 (20.54–35.72)
NH Asian or Pacific Islander
3046 (3.74%)
9.27 (7.65–10.96)
NH Black or African American
6846 (8.41%)
8.36 (7.41–9.31)
NH White
70,206 (86.2%)
9.38 (9.03–9.71)
Hispanic or Latino
14,225 (17.47%)
21.82 (20.02–23.58)
Age
b
65–74 years
50,248 (52.41%)
10.3 (9.22–10.04)
75–84 years
35,207 (36.72%)
11.66 (11.09–12.25)
85 + years
10,415 (10.86%)
8.55 (7.77–9.33)
Urbanization
a
Metropolitan
19,738 (20.59%)
9.94 (9.61–10.27)
Nonmetropolitan
76,132 (79.41%)
11.57 (10.79–12.32)
Place of Death
c
Medical Facility
41,115 (41.85%)
–
Decedent’s Home
29,134 (29.65%)
–
Hospice Facility
4957 (5.05%)
–
Nursing Home/Long-term Care Facility
17,208 (17.51%)
–
Others
5836 (5.94%)
–
Fig. 1.
Overall and Sex-Stratified DM-Related AAMRs per 100,000 in All-LD Patients Among the population of the United States, 1999 to 2020
Diabetes mellitus and all liver disease-related yearly pattern stratified by sex
Men accounted for 50,881 deaths (53.07%) and women for 44,989 (46.93%). The male AAMR was 12.4 (95% CI, 11.91–12.95) versus 8.51 (95% CI, 8.15–8.90) in women. Male rates rose from 10.7 in 1999 to 19.2 in 2020; female rates rose from 8.3 to 12.3. Joinpoint analysis showed minimal change in men through 2012 (APC 0.09%), followed by acceleration during 2012–2018 (4.62%; p = 0.020) and 2018–2020 (12.96%; p < 0.001). In women, rates were flat to slightly declining through 2014 (-0.49%), then increased during 2014–2018 (4.04%) and sharply during 2018–2020 (14.29%; p < 0.001) (Fig. 1 Supplementary Fig. 1, Supplementary Tables 1,3 and 4).
Diabetes mellitus and all liver disease -related yearly pattern stratified by race
Regarding race and ethnicity, across the period, deaths were: The mortality data shows that Non-Hispanic White individuals accounted for 70,206 deaths (86.2%), followed by Non-Hispanic Black individuals with 6,846 deaths (8.4%), Non-Hispanic Asian/Pacific Islanders with 3,046 deaths (3.7%), Non-Hispanic American Indian/Alaska Natives with 1,350 deaths (1.7%), and Hispanic individuals with 14,225 deaths (17.5%). Corresponding AAMRs were highest in AI/AN (27.36; 95% CI, 20.54–35.72) and Hispanic (21.82; 95% CI, 20.02–23.58) groups, compared with NH White 9.38 (95% CI, 9.03–9.71), NH Black 8.36 (95% CI, 7.41–9.31), and NH Asian/Pacific Islander 9.27 (95% CI, 7.65–10.96). Temporal patterns differed: NH White rates climbed from 8.6 in 1999 to 14.3 in 2020, with APC segments of -0.32% (1999–2012), 4.26% (2012–2018; p = 0.0068), and 14.00% (2018–2020; p < 0.001). NH Black rates changed modestly, 9.1 to 10.8, with a recent rise (APC 6.25% during 2015–2020; p = 0.034). Asian/Pacific Islander rates declined through 2016 (APC − 2.14%; p = 0.0088; nadir 7.2 in 2016) and then rose (7.25% during 2016–2020; p = 0.028) to 10.9 in 2020. Hispanic rates were persistently high and stable through 2018 (APC 0.04%), followed by a sharp increase to 29.9 in 2020 (APC 16.04% during 2018–2020; p < 0.001). AI/AN rates were high throughout, increasing from 18.4 in 1999 to 43.7 in 2020, with a pronounced late-period rise (APC 25.19% during 2018–2020; p = 0.0004) (Fig. 2, Supplementary Fig. 2, Supplementary Tables 3 and 5). It should be noted that estimates for smaller racial groups, particularly American Indian/Alaska Native populations, may be influenced by wider confidence intervals and potential misclassification in death certificate data, which could affect the precision of AAMR estimates.
Fig. 2.
DM-related AAMRs in All-LD patients per 100,000 Stratified by Race in the United States, 1999 to 2020
Diabetes mellitus and all liver disease -related yearly pattern stratified by geographical region
In terms of Census regions, the South contributed the largest share of deaths (35,925; 37.47%), followed by the West (24,565; 25.62%), Midwest (20,141; 21.01%), and Northeast (15,239; 15.90%). Cumulative AAMRs mirrored this gradient: West 12.23 (95% CI, 11.48–12.96), South 10.22 (9.71–10.73), Midwest 9.70 (9.04–10.31), and Northeast 8.55 (7.90–9.19). Region-specific rates rose late in the series, reaching in 2020: Northeast 10.9, Midwest 14.6, South 16.7, and West 17.4 per 100,000. APCs showed limited change pre-2010s but consistent acceleration thereafter, with significant late-period increases in all regions—e.g., Midwest 7.26% (2014–2020), South 16.14% (2018–2020), West 7.17% (2015–2020), and Northeast 13.76% (2018–2020) (all p < 0.001 for these segments) (Fig. 3, Supplementary Fig. 3, Supplementary Tables 3 and 6).
Fig. 3.
DM-Related AAMRs per 100,000 in All-LD patients Stratified by Census Region in the United States, 1999 to 2020
State AAMRs varied more than threefold. Among the highest were Texas 16.4 (95% CI, 16.1–16.7), Oklahoma 16.3 (15.6–17.0), West Virginia 16.3 (15.4–17.3), Vermont 15.7 (14.0–17.4), Kentucky 15.5 (14.8–16.1), New Mexico 15.0 (14.0–15.9), and California 14.5 (14.3–14.7). Intermediate-high states included Rhode Island 13.4 (12.2–14.6), Oregon 12.7 (12.1–13.3), Washington 12.4 (11.9–12.9), and Ohio 12.4 (12.0–12.8). The lowest rates were observed in Nevada 5.4 (4.9–6.0), District of Columbia 6.6 (5.4–7.9), Florida 6.7 (6.5–6.9), Louisiana 6.8 (6.3–7.2), Georgia 6.9 (6.6–7.3), and New York 7.0 (6.7–7.2), with several New England and Mid-Atlantic states also on the lower end (e.g., Massachusetts 7.3; Connecticut 7.6; Missouri 7.9; Alabama 8.2) (Fig. 4, Supplementary Table 7).
Fig. 4.
Age Adjusted Mortality Rates in States of U.S. (1999-2020)
Despite most deaths occurring in nonmetropolitan areas (76,132; 79.41%), the urban population experienced lower cumulative risk (urban AAMR 9.94; 95% CI, 9.61–10.27) than rural areas (11.57; 10.79–12.32). Both settings saw marked late-period acceleration: urban APC shifted from − 0.35% (1999–2012) to 3.13% (2012–2018; p = 0.022) and 14.12% (2018–2020; p < 0.001), while rural APC increased from 0.79% (1999–2014; p = 0.008) to 9.44% (2014–2020; p < 0.001). In absolute terms, AAMRs rose from 9.3 to 14.6 in urban areas and from 9.2 to 19.0 in rural areas between 1999 and 2020. (Fig. 5, Supplementary Fig. 4, Supplementary Table 8).
Fig. 5.
DM-Related AAMRs per 100,000 in All-LD patients Stratified by Urbanization among older adults in the United States, 1999 to 2020
Separate analysis of all-LD and DM
Based on the analysis of the extracted data for adults aged 65 and older from 1999 to 2020, mortality trends demonstrate distinct trajectories for diabetes mellitus (DM) and liver disease when analyzed independently, versus the synergistic burden observed in co-occurring cases. For diabetes mellitus alone, the age-adjusted mortality rate (AAMR) decreased significantly from 480.0 per 100,000 population in 1999 to 376.1 per 100,000 in 2020. This represents an overall decline of approximately 21.6% over the study period. In contrast, mortality associated with liver disease alone showed a divergent trend, with the AAMR increasing from 89.8 per 100,000 in 1999 to 110.1 per 100,000 in 2020, reflecting an overall increase of roughly 22.6% (Supplementary Tables 9 and 10).
These findings highlight a critical disparity in the mortality burden of these conditions. While public health interventions and management strategies appear to have reduced overall mortality associated with diabetes in older adults, the burden of liver disease has progressively worsened. This divergence in trends may suggest an increasing relative contribution of liver disease to mortality in this population. However, these observations should be interpreted with caution, as they are not based on formal competing risk analyses and may also be influenced by changes in death certification practices over time. The co-occurrence of these conditions remains associated with a substantial and growing mortality burden, as reflected in the primary analysis of comorbid diabetes and liver disease.
Discussion
We report a comprehensive 21-year analysis (1999 to 2020) of liver disease and diabetes mellitus (DM)-related mortality using the CDC-WONDER database. This study was stratified by sex, race, region, and urban-rural classification, and highlights multiple significant demographic and geographic disparities. Age-adjusted mortality rates (AAMRs) across the years reveal concerning trends, especially in recent years, with clear associations to broader clinical, metabolic, and socioeconomic determinants [18, 19]. Throughout the 21-year period, AAMR rose from 9.3 in 1999 to 10.2 in 2020, reflecting the increasing clinical and public health burden of comorbid liver disease and DM. The mortality trend remained relatively stable in the early years with a nonsignificant change from 1999 to 2014 (APC: 0.08%, p = 0.722). A notable increase emerged from 2014 to 2018 (APC: 5.47%). This upward trend began prior to the pandemic, potentially reflecting the rising prevalence of metabolic dysfunction-associated steatotic liver disease (MASLD) in the aging population. The trend then accelerated sharply from 2018 to 2020 (APC: 12.42%, p < 0.000001), a period encompassing the COVID-19 pandemic. (Central Illustration).
Sex-based stratification showed a higher AAMR consistently among males. Men had a peak AAMR of 19.2 per 100,000 in 2020, compared to 12.3 in women. While men showed steady increases since 2012 (APC: 4.62%, p = 0.019), the rise accelerated in 2018–2020 (APC: 12.96%, p < 0.000001). The propensity to have lower rates in females could be attributed to the protective effects of estrogen against hepatic fibrosis and insulin resistance whereas in males, increased alcohol exposure, later healthcare-seeking behaviour, and occupational exposures may drive the male predominance in deaths of cirrhosis-DM [20–22].
Evidence of racial disparities also existed. AI/AN individuals were always the group with the highest AAMRs, which in 2020 equated to 43.7. This was followed by Hispanics with an AAMR of 29.9. Asian or Pacific Islanders had the lowest overall AAMR, though they experienced a significant rise from 2016 to 2020 (APC: 7.25) [19, 23]. The high rates experienced by the AI/AN community indicate the underlying structural disparities like poverty, geographic isolation, and fewer opportunities to care (only 63.1% report information on access to providers). Given the high average age of this cohort (≥ 65), these structural barriers likely compounded chronic disease progression over decades. Moreover, the prevalence of obesity, sugary drinks, and metabolic syndrome are also more common in these groups [24]. The elevated rates in the Hispanic population could be connected to PNPLA3 gene variant that predisposes steatosis and fibrosis, adding to effects of DM on liver pathology [19].
There is also a widening over time in urban-rural disparities [25, 26]. Rural areas exhibited higher overall AAMRs (11.57 versus 9.94), with mortality increasing strongly after 2014 (APC: 9.44, p < 0.000001). The possible contributing factors are lower access to specialists, limited healthcare resources, and socioeconomic disadvantage along with healthcare reluctance and cultural views [20, 21]. Furthermore, rural ‘medical deserts’ often lack specialists such as endocrinologists and hepatologists, leading to delayed management of complications. In addition, delayed diagnosis and lack of regular screenings in rural regions increase the risk of decompensation. Limited telehealth infrastructure and shortage of healthcare personnel in rural counties may have worsened outcomes, particularly during the COVID-19 crisis, where access to timely interventions was disrupted [27].
Census regional analysis showed the West had the highest cumulative AAMR (12.23), followed by the South (10.22). The Northeast had the lowest rate (8.55). States such as Texas (16.4), Oklahoma (16.3), and West Virginia (16.3) recorded the highest mortality rates, consistent with elevated diabetes, obesity, and alcohol use burdens. Conversely, states like Georgia (6.9) and Florida (6.7) had among the lowest. California, despite its large population, showed a high AAMR (14.5), possibly due to large Hispanic and AI/AN populations at higher risk. These variations point to local policy differences, healthcare infrastructure, and sociodemographic makeup [23, 28]. For example, variations in healthcare infrastructure, such as lower Medicaid expansion or higher uninsured rates, may drive higher mortality due to delayed intervention for chronic comorbidities.
Place of death trends reflected shifting healthcare use patterns. From 1999 to 2020, home deaths increased substantially (from 724 to 3321), while deaths in medical facilities remained relatively stable. This may reflect end-of-life preferences, hospice expansion, or healthcare system strain during COVID-19 [29, 30]. Increased comfort with dying at home, cultural preferences, and limited ICU capacity during pandemic surges could also explain the shift [27, 31].
The significant rise from 2018 to 2020 aligns with the disruptive impact of the COVID-19 pandemic, contributing both directly and indirectly. Lockdowns, reduced outpatient visits, and healthcare resource reallocation delayed care for chronic diseases, while viral infection may have precipitated decompensation in cirrhotic diabetics. Moreover, emerging evidence suggests COVID-19 can directly worsen liver injury and disrupt glucose metabolism, potentially accelerating complications in these vulnerable populations [32, 33].
These trends likely reflect the rising burden of NAFLD and metabolic syndrome with age. The elderly are especially vulnerable due to declining estrogen levels (in women), beta-cell exhaustion, and coexisting cardiovascular diseases. Importantly, CVD is a leading cause of death in individuals with DM and NAFLD, further compounding mortality. Polypharmacy, frailty, and delayed liver disease diagnosis in geriatric groups also add to the complexity of management [21, 34].
Ultimately, the data indicate a worsening burden of coexistent liver disease and diabetes, driven by aging populations, obesity epidemics, and healthcare inequities. These findings underscore the urgent need for targeted screening, public health interventions, and equity-focused policy to mitigate future mortality trends. Policy measures should prioritize increased access to early screening, culturally tailored health education, and preventive metabolic care, especially in under-resourced and high-risk communities. Addressing food insecurity, sedentary lifestyles, and insurance gaps will be essential in curbing this dual epidemic. Furthermore, research efforts must explore better biomarkers and predictive tools to identify those at highest risk for liver-related complications among diabetics, ensuring timely intervention and improved survival outcomes [35].
Lastly, our supplementary analysis of single-disease mortality reveals a striking divergence that illuminates the drivers of our primary findings. Nationally, mortality rates for Diabetes Mellitus alone decreased by 21.6% between 1999 and 2020, reflecting decades of improvements in cardiovascular risk management and glycemic control. In sharp contrast, mortality for All Liver Diseases alone rose by 22.6% over the same period. The fact that our study cohort experienced a significant surge in mortality suggests that the rising burden of liver disease is overpowering the survival gains typically seen in diabetes care. This ‘competing risk’ phenomenon indicates that as older adults survive longer with diabetes due to better cardiac care, they are increasingly succumbing to hepatic complications. The synergistic rise likely stems from the untreated progression of metabolic dysfunction-associated steatotic liver disease (MASLD) and the compounding impact of insulin resistance on hepatic fibrosis, which creates a frailer phenotype highly susceptible to acute decompensation [36, 37].
Limitations
While this study offers valuable insight into national trends in All-LD and DM-related mortality, several limitations must be acknowledged. First, reliance on death certificate data means we analyzed reported causes of death rather than clinically confirmed prevalence. While we utilized multiple cause of death codes to capture the burden of these conditions as contributing factors, we cannot confirm the exact clinical sequence of events leading to death. Second, the dataset lacks clinical granularity regarding disease stage, treatment regimens, or glycemic and hepatic control. Third, social determinants of health such as income, education, healthcare access, and housing instability were not captured but likely contribute significantly to the observed disparities. Additionally, the impact of alcohol use versus non-alcoholic liver disease could not be fully disentangled. Mortality estimates derived from national databases may be subject to misclassification of race and ethnicity on death certificates, particularly among American Indian/Alaska Native populations, where underreporting has been previously documented. Additionally, smaller population sizes in certain racial groups may result in less stable AAMR estimates, reflected by wider confidence intervals. These factors should be considered when interpreting disparities across racial and ethnic groups. Future research should incorporate longitudinal clinical data and consider sociodemographic and behavioral risk factors to better understand the drivers of mortality and inform tailored interventions.
