INTRODUCTION
Although global life expectancy continues to rise, older adults’ remaining years do not guarantee good health or autonomy. According to a recent Eurostat report, more than half of those aged 65 and older spend much of their remaining years with functional limitations, moderate disability, or severe activity limitations [1]. Rising disability in older adults is a major factor in reducing independence, worsening depression [2], and lowering quality of life [3]. Many older adults with severe functional dependence live in long-term care facilities (LTCFs) [4], placing a huge burden on society by increasing care and medical expenses [1].
In Korea, LTCFs operate under two systems: Long-Term Care Insurance and National Health Insurance (NHI), which support nursing homes (NHs) and long-term care hospitals (LTCHs), respectively [5]. NHs provide assistance with daily living, whereas LTCHs provide long-term inpatient care centered on medical and nursing services. Both facilities play a key role in managing functional decline and impairment of activities of daily living (ADL) among long-term care recipients. Recently, the Ministry of Health and Welfare has promoted integration of the two systems through the Integrated Medical and Care Support Pilot Project and the Integrated Community Care Assistance Act, aiming to improve a service system previously divided by facility type [5].
ADLs are basic self-care activities performed independently and serve as a key indicator of older adults’ independence [1,3]. Previous studies identify multiple factors related to ADL decline in older adults, including demographics, socioeconomic and psychosocial factors, health behaviors, health status, and physical function. Older age, female sex, disease, cognitive decline, and physical decline consistently contribute to decreased ADL [6]. Although Korea’s long-term care system is institutionally divided between NHs and LTCHs, frequently transfers occur between facilities as older adults’ health status changes [5].
Most studies in Korea have focused on either NH or LTCH, with few analyzing both together. In particular, few studies have compared health- and care-related determinants of ADL disability across age groups. Additionally, since severe ADL disabilities (S-ADLD) are directly correlated with increased mortality [7], further research is needed to identify age-specific patterns and factors of S-ADLD, considering the types of LTCFs.
To better represent older adults in Korea, this study defined residents of ‘LTCFs’ as those living in either NHs or LTCHs. This study aimed to identify the prevalence of ADL disability and major age-specific risk factors (young-old, mid-old, and old-old) thereby providing basic data to guide nursing interventions and policy strategies for long-term care.
METHODS
Ethic statements: This study was approved by the Institutional Review Board (IRB) of Hanyang University (IRB no. HYUIRB-2020-200). Written informed consent was obtained from all participants, ensuring their understanding and agreement to participate in the study.
1. Study Design
This cross-sectional analysis used first-year data from a longitudinal study (September 2020 to February 2025) and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (https://www.strobe-statement.org/).
2. Study Participants
Data for this study were collected from the corresponding author’s ongoing longitudinal study on risk factors for mortality among older adults residing in LTCFs. For present analysis, cross-sectional first-year data were used to identify factors related to ADL disability [8].
Participants were adults aged 65 years or older admitted to NHs and LTCHs, defined as residents of LTCFs. Inclusion criteria were older adults aged 65 years or older, ability to communicate, residence in the facility for at least 2 months, and informed consent after understanding the study purpose. For older adults with cognitive impairment, facility staff (nurses) recommended participants who could communicate verbally and nonverbally. Capacity to consent was assessed through simple conversations about greetings, weather, meals, and/or daily life and participation included only those who voluntarily agreed. Of 576 participants recruited, one with missing Barthel Index ADL data and 11 with normal ADL based on the Barthel Index score were excluded (four aged 65~74 years, four aged 75~84 years, and three aged 85 years or older). The final analysis included 564 participants (67 aged 65~74 years, 211 aged 75~84 years, and 286 aged 85 years or older). Sample size was determined using an event per variable standard of 10 or more [9] to ensure regression model stability.
This study strictly adhered to ethical procedures to protect the rights of vulnerable older adults residing in LTCFs. All procedures followed the ethical principles of the Declaration of Helsinki and were registered in the Clinical Research Information Service (KCT0006437). Only residents aged 65 years or older who voluntarily agreed to participate after understanding the study purpose were included. For participants with mild cognitive impairment, the ability to provide informed consent was assessed through brief daily conversations (e.g., greetings, weather, meals). When necessary, family members or legal guardians were provided with detailed explanations and gave written proxy consent. Participants’ willingness to continue participation was reconfirmed throughout data collection, and all were informed that they could withdraw from the study at any time without penalty. Facility participation was approved by each institutional administration, and all data collection procedures complied with infection-control and privacy-protection policies to ensure participants’ safety and well-being.
3. Measurements
1) Dependent Variable
ADL independence was measured with the Barthel ADL Index, developed by Collin et al. [10] and translated into Korean [11]. The Barthel ADL index consists of 10 items (bowel control, urinary control, toilet use, grooming, eating, dressing, transferring, mobility, stair use, and bathing), each scored as ‘0 to 2’ or ‘0 to 3’ points. Some were first calculated on a 20-point scale [10], and then converted to a 100-point scale, where 100 indicated complete independence (i.e., no disability). All participants who had some degree of disability were classified as moderate ADL disability (Barthel Index≥70) or S-ADLD (Barthel Index<70), based on the Welfare of Persons with Disabilities Act in Korea. For analysis, S-ADLD was dichotomized as 1=yes and 0=no. Cronbach’s alpha was .94 for Korean version [11] and .93 in this study.
2) Independent Variables
Sex was recorded as 0 for males and 1 for females. Excessive polypharmacy was coded as 0 for nine or fewer prescribed drugs and 1 for 10 or more drugs. This was based on a previous study [12] that defined excessive polypharmacy as taking 10 or more drugs simultaneously. Incontinence was coded as 0 indicating neither and 1 indicating urinary or fecal incontinence. Dietary changes were measured with the Korean version of Mini Nutritional Assessment–Short Form (MNA-SF) provided by the Nestlé Nutrition Institute [13,14]. Appetite loss was assessed with the question: “In the past 3 months, has your appetite decreased due to loss of appetite, indigestion, or difficulty chewing or swallowing?” Responses “a lot decreased,” “a little decreased,” or “no change” were coded as 0 for no change, and 1 for “a little decreased” and “a lot decreased.” Calf circumference (CC) was measured as an indicator of nutrition and frailty. Low CC was coded as 0 for those in the 25th percentile or higher of the participant’s quartile and 1 for those lower than 25% [15]. The lower 25% criterion was based on 25.0 cm or less for women and 28.0 cm or less for men. Cognitive function was measured using the Cognitive Impairment Screening Test (CIST) [16], which includes 13 items across six domains: orientation, memory, attention, visuospatial function, language ability, verbal ability, and executive function. Items were scored 0~1 or 0~2, with a total score of 30; higher scores indicated better function. The assessment is differentiated by norms based on age (years in Korean) and educational level. Above the normal standard was scored as 0, and below the normal standard was scored as 1. In this study, the Cronbach’s α of the CIST was .90. LTCFs were coded 0 for NHs and 1 for LTCHs.
3) Stratification Variable
This study conducted an age-stratified subgroup analysis to examine the differences in the effects of the independent variables across age groups. Participants were stratified into three age groups: young-old (ages 65~74), mid-old (ages 75~84), and old-old (ages 85 and older), based on criteria commonly used in gerontology.
4. Data Collection
Data collection for this study was conducted after obtaining approval from the Institutional Review Board (IRB) of Hanyang University (IRB No. HYUIRB-2020-200). This study analyzed factors related to ADL disability in older adults in LTCFs, differing from a prior analysis of frailty and nutrition using the same first-year data [8]. The study was conducted in 17 LTCFs (NHs and LTCHs) located in seven regions of South Korea (Seoul, Gyeonggi, Incheon, Daejeon, Chungnam, Chungbuk, and Gyeongbuk). Participating facilities were selected to reflect variation in size (50~200 beds) and type (medical-based or social welfare-based). Due to COVID-19 restrictions, trained facility staff (nurses, nurse aides, and social workers) who had completed a standardized online education program (seven modules, approximately 120 minutes) administered the survey directly. The research team provided electronic manuals and instructional videos to ensure standardization of data collection procedures. All data collection processes were monitored in real time by the research team through an online system, and immediate feedback was provided to minimize missing data and maintain data consistency and reliability. For example, CC was measured using a plastic tape while participant’s knees were bent at 90°, recording the average of both calves. To minimize interrater error, a manual and online counseling support were also provided. The data collection procedure was previously described in detail in a previous study [8].
5. Data Analysis
Participant characteristics were summarized as frequencies and percentages, and/or mean±standard deviation. Differences between age groups in characteristics and S-ADLD were tested using the chi-square test; Fisher’s exact test was applied when expected frequencies were less than five. Risk factors for S-ADLD by age group were examined with multivariate logistic regression, including significant univariate variables and seven independent variables reported in previous studies [6,17,18]. Analyses were conducted in IBM SPSS Statistics version 28.0 (IBM Corp.). Statistical significance was set at p-value less than .05. Forest plots of odds ratios (ORs) with 95% confidence intervals (CIs) were generated in Python 3.11 using matplotlib and seaborn. Missing values were observed for polypharmacy (0.4%), incontinence (0.9%), CC (1.6%), and cognitive function (8.0%), as all were less than 10%, no imputation was performed [19].
RESULTS
1. Severe ADL Disability
Of the total participants, 65.4% had S-ADLD. By age group, S-ADLD prevalence was 64.2%, 67.3%, and 64.3% in the young-old, mid-old and old-old groups, respectively. The distribution of S-ADLD across age groups did not differ significantly (Table 1).
2. Geriatric Characteristics in Older Adults by Age Group
Characteristics by age group are presented in Table 1. The mean age of all participants was 83.58±7.42 years; by age group, mean ages were 70.08±2.97 years for young-old, 80.15±2.59 years for mid-old, 89.46±3.80 years for old-old. The geriatric characteristics differed by age group only for sex and long-term care setting. Overall, 78.9% of participants were female. By age group, males were most prevalent at 52.2% in the young-old group, whereas females were significantly more prevalent in the mid-old (79.1%) and old-old group (86.0%) (p<.001). Overall, 68.1% of the participants were in NHs, compared with 31.9% in LTCHs. Among those aged 85 years or older, the proportion in NH was the highest at 74.5% (p<.001). The remaining variables (excessive polypharmacy, any type of incontinence, dietary changes, low CC, and suspected cognitive impairment) showed no significant differences among age groups.
3. Differences in Severe ADL Disability According to Characteristics of Older Adults by Age Group
Relationships between age group characteristics and S-ADLD are shown in Table 2. S-ADLD was significantly associated with any type of incontinence, dietary changes, low CC, suspected cognitive impairment, and LTCF residence. Although any type of incontinence and low CC were consistently associated with S-ADLD across all age groups, sex and excessive polypharmacy were not. Specifically, the rate of S-ADLD was high among participants with any type of incontinence (p<.001), and the difference was significant in all age groups: young-old (p=.041), mid-old (p<.001), and old-old (p<.001). Dietary change was associated with a higher rate of S-ADLD overall (p=.005) and was significant only in the mid-old age group (p=.001). Low CC (lowest 25%) was significantly associated with S-ADLD (p <.001) and in all age groups. Among participants with suspected cognitive impairment, S-ADLD was more prevalent (p<.001); by age group, this association was significant only in the mid-old (p=.001) and old-old age group (p=.025). In addition, among those old-old age group, the rate of S-ADLD was significantly higher in LTCHs than in NHs (p<.001).
4. Factors Related to Severe ADL Disability
The results of the logistic regression analysis to determine the risk factors for S-ADLD are presented in Figure 1. In the overall participants analysis, any type of incontinence (OR=3.20, 95% CI = 2.10~4.87, p<.001), dietary change (OR=1.87, 95% CI=1.02~3.45, p=.043), low CC (OR=4.44, 95% CI=2.41~8.19, p<.001), suspected cognitive impairment (OR=2.42, 95% CI=1.42~4.11, p=.001), and LTCH residence (OR=2.65, 95% CI=1.63~4.32, p=.001) were significantly associated with S-ADLD. Specifically, in the young-old age group, female (OR=4.18, 95% CI=1.04~16.83, p=.044) had a significantly higher risk of S-ADLD than male. Low CC (OR=5.90, 95% CI=0.90~38.78, p=.065), although not statistically significant, showed a relatively high OR, suggesting a potential association with the risk of S-ADLD. In the mid-old age group, any type of incontinence (OR=3.22, 95% CI=1.58~6.55, p=.001), dietary change (OR=4.74, 95% CI=1.47~15.26, p=.009), low CC (OR=3.09, 95% CI=1.10~8.66, p=.032), suspected cognitive impairment (OR=3.27, 95% CI=1.35~7.96, p=.009), and LTCH residence (OR=2.41, 95% CI=1.09~5.36, p=.031) were significantly associated with S-ADLD. In the old-old group, any type of incontinence (OR=3.87, 95% CI=2.13~7.05, p<.001), low CC (OR=6.38, 95% CI=2.65~15.34, p<.001), suspected cognitive impairment (OR=2.24, 95% CI=1.02~4.94, p=.045), and LTCH residence (OR=3.04, 95% CI=1.42~6.54, p=.004) were significantly associated with S-ADLD.
DISCUSSION
In Korea, national longitudinal data on older adults living in the community have been established, and research is being actively conducted. However, there is little systematic data on those residing in LTCFs. This study is significant because it analyzed the health characteristics and functional decline patterns of older adults in LTCFs based on multidimensional data, including physical measurements. Among participants (mean age, 83.58 years), all had ADL impairment, and 65.4% had S-ADLD.
When older adults were analyzed together, incontinence, dietary changes, low CC, suspected cognitive impairment, and LTCH residence were the main predictors of S-ADL. When older adults were divided by age group, the differences were clear. In the young-old age group, only sex was a significant factor, but in the mid-old age group, various factors such as urinary and fecal incontinence, dietary changes, decreased CC, suspected cognitive decline, and LTCHs were involved. In contrast, in the old-old group, the significance of dietary changes disappeared, and urinary and fecal incontinence, decreased CC, and LTCHs remained significant factors. This suggests that the detailed factors necessary for an actual intervention may be overlooked when interpreted as a whole. Therefore, analysis by age group is a strength of this study because it enables the establishment of customized intervention strategies based on characteristics. Thus, we describe the main results according to the age group.
The fact that only sex (female) was significant in the young-old group suggests that the overall physical function and health status were relatively maintained in this age group. This is unlike the accumulation of various risk factors in the mid-old group. Compared to male, female experience accelerated muscle loss and musculoskeletal vulnerability due to a decrease in estrogen after menopause [20], and the risk of muscle weakness and falls increases. These characteristics appear to be major factors in S-ADLD. In contrast, decreased CC and LTCHs were not statistically significant in the young-old group, but the ORs were high. Therefore, the potential risk of S-ADLD associated with early sarcopenia and LTCH residence cannot be excluded. Early screening and preventive nursing interventions focusing on sex-related factors are especially important in the young-old group, distinguishing them from the complex risk factors that appear in later age groups.
In the mid-old group, any incontinence, dietary change, low CC, suspected cognitive impairment, and LTCHs were all identified as the most common and significant risk factors for S-ADLD. This age group is a transitional period in which physical, cognitive, and nutritional impairments accumulate and is characterized by simultaneous multifactorial risk factors. Dietary changes and incontinence are easily overlooked early signs of functional decline, although these are the most observable daily indicators in clinical practice. In this study, decreased appetite was significantly associated with S-ADLD, and this trend occurred only in the mid-old group, suggesting that different factors depend on age. Poor appetite is a common problem in older adults that causes a decrease in food intake, ultimately leading to malnutrition [21], frailty [22], and even death [21]. This is consistent with previous studies conducted in domestic LTCFs, where malnutrition was identified as a major factor related to frailty [8]. Therefore, there is a need to improve standardized screening, evaluation, and management systems for poor appetite in older adults residing in LTCFs.
Any type of incontinence has been confirmed as a consistent predictor of S-ADLD in all age groups over 75 years of age, and this trend aligns with the previous studies [4,23]. In a study of older adults residing in LTCFs in Spain, the prevalence of urinary incontinence among people aged 65 or older living in NHs was high (76.5 %), and urinary incontinence was independently related to ADL limitation [23]. In addition, a previous study conducted on people with cognitive impairment living in LTCFs in Poland [4] reported that urinary and fecal incontinence are common health problems among older adults in NHs.
CC was a significant risk factor for S-ADLD in all age groups over 75 years. Previous research also reported that CC measurement is a simple indicator that reliably reflects a decreased muscle mass in older adults residing in LTCFs [24,25], and is an effective tool for screening sarcopenia [26]. Sarcopenia is a major geriatric syndrome that causes various health problems in older adults and is particularly common in NHs [27]. A 15-year tracking of CC in older people confirmed that CC decreases, and functional deterioration progresses with age [28]. This suggests that CC is an important indicator of S-ADLD in older adults residing in NHs after midlife.
Cognitive decline is also consistently identified as predictive factor for S-ADLD in all age groups over 75 years of age. In a previous study on a long-term care cohort, cognitive and functional decline were reported to be very common in older adults in NHs, and cognitive impairment was reported to have a major impact on ADL disability by impairing planning and executive functions [29].
LTCH residence was a consistent predictor of S-ADLD in all age groups over 75 years of age. In particular, the S-ADLD rate of people residing in LTCHs was higher than that of people residing in NHs, which is the same result as in a previous domestic study [30] and reflects the structural characteristics of domestic LTCFs that mainly accommodate older adults with high medical severity. In Korea’s long-term care system, the proportion of people admitted to LTCHs is very high, which is closely related to the fact that LTCHs appear to be the most common place of death [5]. This structure reflects the reality, where people with high functional dependence are concentrated in LTCHs. This may lead to a high incidence of S-ADLD in LTCHs. Thus, it is necessary to introduce customized management and rehabilitation programs that consider differences in the types of LTCFs. The appearance of multiple risk factors identified in the mid-old population is an important indicator of a transitional period of poor health and may be the optimal time for interventions to prevent functional decline.
In this study, the risk of S-ADLD was significantly higher among older adults residing in LTCHs. This finding reflects the greater functional dependence and higher medical severity of LTCH residents and indicates structural differences between Korea’s two long-term care systems: LTCHs are covered by NHI, whereas NHs are supported by the Long-Term Care Insurance. Consequently, resident characteristics and health conditions differ substantially between the two facility types. LTCHs mainly accommodate older adults requiring continuous medical treatment and nursing care, whereas NHs primarily provide daily living assistance. Therefore, the significant association between LTCH residence and S-ADLD in this study likely reflects these institutional and demographic differences. However, as only facility type was included as a variable, detailed individual-level clinical indicators (e.g., number of comorbidities, past medical history, muscle strength, depression level, and use of medical devices) were not adjusted for. Future studies should incorporate these indicators to provide a more comprehensive and precise understanding of residents’ health status and the differences between facility types.
In the old-old group, urinary incontinence, nutritional status, low CC, decreased cognitive function, and residence in LTCHs acted as universal and persistent risk factors for S-ADLD, while dietary changes did not show a significant relationship. This can be interpreted as a ‘floor effect’ that occurs as chronic loss of appetite exists in most of the very old age groups. A previous study [21] also reported that decreased appetite is closely related to poor prognosis, such as malnutrition and death, in older adults in all environments, including the community, NHs, and hospitals. However, in this study, it can be inferred that the discriminatory power as a predictor was weakened because decreased appetite was already widespread in the old-old group, suggesting that the effect of appetite on S-ADLD may vary depending on age. As a result, the weakening of dietary indicators in the old-old group in this study suggests that appetite may no longer be a key determinant of S-ADLD and that intervention strategies for this age group need to focus on core physiological functions such as muscle maintenance and urinary function rather than nutrition.
CONCLUSION
This study has great academic and policy significance because it identified functional decline and risk factors in high-risk groups based on the actual measurement data of people living in LTCFs, including both NHs and LTCHs, which has rarely been reported in Korea. Continuous accumulation of such data will strengthen the basis for long-term care policies and contribute to the development of care systems in an aged society. The results also show the need for differentiated management strategies according to the age group. In other words, the young-old group requires muscle strengthening, skeletal health promotion, fall prevention, and exercise programs for female that consider their sex characteristics. For the mid-old group, integrated management that simultaneously addresses urinary incontinence, nutritional imbalance, muscle loss, and cognitive decline was effective. As the risk to residents in LTCHs increases, customized strategies for each facility type are needed. For the old-old group, a strategy that focuses on maintaining core physiological functions, such as maintaining muscle mass, managing urinary incontinence, and preventing cognitive decline, rather than poor appetite, will be effective.
This study has several strengths in identifying risk factors for ADL disability among people residing in LTCFs. First, the impact on S-ADLD disability was confirmed comprehensively and specifically through multivariate analysis, including physical measurements and facility type (NHs and LTCHs). This is significant because it reflects the situation in which older adults use the two facilities in Korea’s long-term care reality, thereby increasing the real-world applicability and external validity of the analysis. Second, by subdividing age groups (65~74 years, 75~84 years, 85 years or older) and comparing the risk factors by age group, factors that act differently at each stage of aging were identified. This approach avoids treating older adults as a single group and provides a basis for customized nursing interventions and policy design for each stage of life. However, this study had an insufficient sample size in some age groups and was based on a cross-sectional design, which limits causal interpretation. Additionally, facility-level variables, such as the ratio of older adults to nursing staff, could not be considered. Therefore, future research should secure a sufficient sample, conduct an analysis by facility type, and use longitudinal data to identify causal relationships.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
