The Impact of Sustainable Development on the Public Health System of the Elderly in the Internet of Things Environment

Abstract

In order to explore how to establish an effective public health and long-term health care system for the elderly in the Internet of Things environment, sustainable development of the public health system for the elderly in the Internet of Things environment was proposed. Through field research, the need of the elderly for home-based care can be divided into the following four categories: daily care, medical care, emergency care, and emotional care. Through the key technical problems and solutions of information recommendation represented by the Internet of Things, we explored the research on how to establish the elderly care system. Existing research shows that nearly 90% of the elderly chose their own home or their children’s to provide care for the elderly, while the proportion of rural elderly who chose family members to provide care for the elderly reaches 97%. However, as an important supplement to family support, the acceptance of institutional support in the public is not optimistic; among them, the overall proportion of the elderly who favored nursing homes as providers for elderly care in the future was less than 5%, and the willingness and attitude of the rural elderly regarding the institutional elderly care model were more disapproving, with only 1.44% of them recognized. Therefore, sustainable development in the Internet of Things environment is not only a key measure to effectively solve the mismatch between supply and demand of elderly care services, but also a favorable exploration for information technology to play an important role in the field of public services.

1. Introduction

The development of medical and health care and the sustainable development of society are in dialectical unity, and there is a two-way interaction between them. The sustainable development of medical and health care is required to solve a series of deep-seated complications. The sustainable development of medical and health care should be defined as the rational allocation and utilization of health resources—which can not only meet the immediate needs of the people—and also contribute to the realization of long-term human health goals. Technology has entered every industry including the medical, commercial, and financial sectors. Medical care can be said to be one of the fastest areas to implement the technological change of the Internet of Things, mainly in terms of thorough diagnosis and treatment. With the rapid development of the social economy and the continuous improvement of people’s living standards, the average life expectancy of human beings is constantly lengthening; conversely, it has brought a severe challenge to the problem of providing care for the elderly all over the world [1]. According to preliminary statistics, the number of disabled elderly people is 33 million, approaching 100 million in this century. The rapidly growing elderly population combined with a system that does not adapt to economic developments not only threaten the endowment and medical security of the elderly, but also highlights the problem of elderly care. As regional economic differences in China have brought about the fragmentation of care in rural areas, a large number of young and middle-aged workers have gone out to work, and the effectiveness of family care is slowly weakening. At the same time, due to the constraints of economic conditions, the development of social care is not perfect, unable to meet the needs of the disabled elderly in rural areas. The problem of care for the disabled elderly has evolved into a huge social risk [2].

In China, the responsibility of care and support for the elderly in their old age is often borne by their adult children, especially with the current trend of fewer children and core family development and their family role of “having both the elderly and the children”, which increases their burden of caring for the elderly. In addition, a large number of the elderly have a moderate or severe disability, which increases the difficulty of care for family care staff. The huge economic and mental pressure creates extreme emotional fluctuations, which leads to many unnecessary family conflicts, affecting the harmony, stability, and development of the family. Therefore, the construction of a multi-level and multi-angle long-term care service system can not only form complementary advantages among different service modes, but also help to ease the conflicts and disputes within the family to promote the spread of good moral culture such as “respect and love the elderly” in society as a whole [3,4].

According to the non-competition of consumption and the non-exclusivity of benefit, products and services can be divided into public goods, quasi-public goods, and private goods. A product that is both non-competitive and non-exclusive is a public product, while a product that is neither non-competitive nor non-exclusive is a private product. Quasi-public goods have only one of two characteristics. The long-term care service for the elderly is a public service that mainly includes the treatment, rehabilitation of various diseases, diagnostic and therapeutic examinations, and corresponding drug consumption. These services are in high demand, expensive, competitive in consumption, and exclusive in benefit, so they fall into the category of quasi-public goods [5].

From 2020, Lukang medical care will provide a combination of medical and nursing services in old-age care, with the combination of medical and nursing as the entry point. It provides services such as life care, basic medical care, aging prevention, geriatric rehabilitation, rehabilitation nursing, psychological comfort, long-term and hospice care (palliative care), and other comprehensive services for the disabled and mentally disabled elderly, the self-care and semi-self-care elderly, disabled persons (including intellectual and mental disability), those with chronic diseases, and hospice patients [6]. The service objects are shown in Figure 1.

On the basis of this research, this paper proposes an impact of sustainable development on the public health system of the elderly in the Internet of Things environment. Through field research, the needs of the elderly for home-based care can be divided into the following four categories: daily care, medical care, emergency care, and emotional care. The disabled elderly are divided into three levels of nursing according to the same standard with a total of ten services, including home service, daycare, family care, home care, home and community rehabilitation services, auxiliary equipment purchase (lease) and home accessibility improvement services, elderly nutrition catering services, transportation services, and long-term care institutions. Finally, big data technology is used to develop a personalized rehabilitation training plan for the elderly in line with their own physical conditions and conduct systematic behavioral guidance for the elderly according to the rehabilitation training plan.

2. Literature Review

Gonzalez, L. et al. said that, in recent years, many scholars have explored individuals’ aversive perception of social inequality based on self-centered fairness theory [7]. The aversion to injustice means that people resist the unfair results of social distribution; that is, they are willing to give up some material rewards in the direction of a more equitable outcome. The heterogeneity of preferences also has an important interaction with the economic environment. The differences in the economic environment in different countries also affect the preference types of the elderly in choosing long-term care services, and these preference types affect the behavior of suppliers in the equilibrium of the long-term care services market. Guo, C. et al. found that the lifetime use rate of assistive devices for the disabled elderly significantly increased from 4.96% in 1987 to 9.07% in 2006 (ptrend < 0.001), and the use rate of medical services in the past 12 months significantly increased from 18.6% in 2007 to 56.9% in 2014 (ptrend < 0.001). What is the impact on public health practice? This arduous achievement shows the success of China’s social, economic and medical reform, and the improvement of health equity. However, more sound policies and actions are needed to further reduce unmet needs in medical services [8]. This indicates that the economic environment of each country/region has an important impact on the construction of long-term care service systems under the existence of demand heterogeneity preference.

On the premise of studying the long-term care insurance system in Germany, Naresh V.S. established public long-term care insurance under the strong support of the government and private long-term care insurance according to the demand of the market in order to cope with the aging population and the changes of social structure [9]. Public long-term care insurance is mandated by the state for everyone. To prevent future care problems, private long-term care insurance depends on an individual’s economic level and personal care needs. Morozova, E. mainly studied the nursing insurance system, a measure taken by Japan to cope with aging, which established a new care prevention system, established comprehensive community assistance centers, and created new care services to meet the care needs of the elderly [10]. From the perspective of the long-term care system in the United States, developed countries mainly guarantee the financing of future social care through the establishment of insurance systems.

Melo, M.T.S.M., in describing social needs (also known as “belonging needs”), pointed out that when physiological and security needs are met to a large extent, people desire to establish an affectionate relationship with ordinary people [11]. The human need for interpersonal attachment and a sense of belonging to others is thought to be fundamental to the species. Kyomba, G.K. believed that the need for belonging is innate and universal because it exists in every human society; at the same time, they also pointed out that social exclusion is likely to be the main cause of people’s anxiety, because it is accompanied by negative feelings such as isolation, loneliness, and depression [12].

Within the research framework of economic theory, Boltz, M. extended the analysis method of demand and supply to the research of non-profit organizations [13]. Sundberg, F. et al. proposed the theory of “government failure”. Its theoretical model is based on the assumption of heterogeneity, that is, the public’s demand for public goods is different, and the higher the heterogeneity is, the more the number and types of NPOs in the region will be [14]. Due to differences in personal income, nationality, religion, and education level, individuals have different needs for public goods. However, the quantity, type, and quality of public goods provided by the government are designed to meet the needs of the majority of people, and the decision-making process also respects the will and interests of the majority of people, with a certain degree of unity. Therefore, there is a contradiction between the heterogeneity of demand and the uniformity of government supply.

3. Methods

Smart elderly care is a sensor network system and information platform for the elderly at home, communities, and elderly care institutions, and on this basis, it provides real-time, fast, efficient, low-cost, Internet of Things, interconnection, and intelligent elderly care services. With the progress of science and technology, new ways of providing for the elderly are becoming more and more popular. A series of high-tech products, such as TV boxes designed only for parents, have emerged in society to improve the quality of life of the elderly in their later years and solve the loneliness problem of the empty nest elderly to the greatest extent. They are new, intelligent forms of providing for the elderly, similar to migratory birds. After more than a year of good operation and rapid growth, Smart Pension has received extensive attention and recognition from the government, industry, public, and media; it allows the elderly to fully enjoy the convenience and comfort brought by the Internet of Things. In 2012, the National Office for the Aging of the People’s Republic of China first proposed the concept of “intelligent elderly care”, encouraging and supporting the practice and exploration of intelligent elderly care. In 2015, the State Council issued the Guiding Opinions on Actively Promoting the Action of “Internet plus”, clearly proposing to “promote the development of smart health pension industry”. In February 2017, the Ministry of Industry and Information Technology, the Ministry of Civil Affairs, and the National Health and Family Planning Commission issued the Action Plan for the Development of Smart Healthy Elderly Care Industry (2017–2020), which plans to build 500 smart healthy elderly care demonstration communities within five years, meaning that smart elderly care is entering the fast lane of development. On 20 April 2022, the General Office of the State Council issued the Opinions on Further Releasing Consumption Potential to Promote the Continuous Recovery of Consumption, pointing out that to adapt to the needs of normalized epidemic prevention and control, promote new consumption, accelerate the organic integration of online and offline consumption, expand the upgrading of information consumption, cultivate and strengthen smart products and smart retail, smart tourism, smart radio and television, smart elderly care, smart housekeeping, digital culture, smart sports, “Internet Plus health care”, “Internet Plus childcare”, “Internet Plus home decoration”, and other new forms of consumption should be adopted.

3.1. Overall Design of Home Care System

Through field investigation and research, the needs of the elderly for home care can be divided into the following four categories: daily care, medical care, emergency care, and emotional care. The specific needs of the elderly are shown in

Table 1. Needs of the elderly for home care.

Needs of the Elderly	Difficulties Faced	The Solution
Daily care	Irregular sleep	To popularize health knowledge, to urge regular rest, personalized life arrangement
	Poor memory	To set event reminders
	Poor ability to accept new knowledge and equipment	To simplify operations for the elderly
Medical care	Do not pay attention to physical function problems, and no physical examination	To push health information, to customize personalized physical examination events
Emergency care	A sudden accident without recourse	Fall detection system, and active alarm system
Emotional care	No communication or lack of communication with children	Message push and care for the elderly
	Passive contact with children	Children end supervision and care for the elderly, children end auxiliary care system

.

3.2. Overall Design of Fall Subsystem

In general, the fall detection system first collects relevant data, and then transmits the collected data to the data processing equipment or unit. The data processing equipment or unit will determine whether the fall behavior occurs on the collected data, and trigger the alarm if the fall condition is met [15]. According to the above workflow, the existing hardware of fall detection equipment consists of the following four parts: first, sensors that collect data such as a three-axis acceleration sensor; second, a communication module, including sending and receiving data; third, the data processing module, whose main function is to preprocess data, remove unnecessary data, select key data and classify them, and then judge the processing results to detect whether there is falling behavior; and lastly, the alarm module, which is responsible for data processing, the alarm module is divided into two parts: warning and formal alarm. The fall detection subsystem of the smart home care system in the study also adopts a similar workflow as above, except that data transmission is omitted, and the existing fall detection algorithm is improved [16]. Firstly, the system collects the user’s acceleration data, preprocesses the original data, and then extracts the features of the input data to form the feature set. Usually, in order to improve the recognition speed of the system, the original feature vector is reduced, and the feature selection process is conducted. This results in an optimized set of features. This series of the feature set is called the sample set, which is sent into the training module, and a classifier can be obtained through a certain training algorithm. The classifier can be defined as shown in Formula (1), and the design of each module of the fall detection device is shown in Figure 2.

$$y=f(x)$$

(1)

where x is the feature vector, y is the output of category results, and f is the representation form that can separate all categories. In this way, when new feature data comes, the classification output can be obtained through certain operations of the classifier.

The built-in acceleration sensor and gyroscope of the iPhone are used as data acquisition tools, and the zero-velocity correction algorithm based on an extended Kalman filter is used to preprocess the data collected by the triaxial acceleration sensor and gyroscope [17]. The acceleration difference estimation vector, angular velocity difference estimation vector, and position difference estimation vector in the iPhone 6 coordinate system are selected as the input vector of the Kalman filter of the iPhone 6. The traditional filtering method can only be achieved when the useful signal and noise have different frequency bands. In the 1940s, N. Wiener and A.H. Kolmogorov introduced the statistical properties of signal and noise into the filtering theory. Under the assumption that both signal and noise are stationary processes, they used optimization methods to estimate the true value of the signal to achieve the filtering purpose. Thus, they are conceptually linked with traditional filtering methods and are called Wiener filtering. This method requires that both signal and noise be stationary processes. In the early 1960s, R.E. Kalman and R.S. Bucy published an important paper, New Results of Linear Filtering and Prediction Theory, and proposed a new linear filtering and prediction theory, which is called Kalman filtering. It is characterized by processing the noisy input and observation signals based on linear state space representation to obtain the system state or real signal. Kalman filter is an algorithm that is applicable to linear, discrete, and finite-dimensional systems. Every autoregressive moving average system (ARMAX) with external variables or a system that can be represented by a rational transfer function can be transformed into a system that can be represented by a state space so that it can be calculated using Kalman filtering. The five Kalman filter formulas in the whole filtering process are shown in Formulas (2)–(6):

$$dxk=F_{k}d{x}_{k-1}+{\omega }_k$$

(2)

$$z_k=Hd{x}_k+n_k$$

(3)

$$\delta x_{k\left|k\right.}:d{x}_{k\left|k\right.}=d{x}_{k\left|k\right.-1}+K_k(z_k-Hd{x}_{k\left|k\right.-1})$$

(4)

$$K_k=P_k\left|{}_{k-1}{H}^T(H{P}_k{}_{\left|k-1\right.}\right.H^T+R_k{)}^{-1}$$

(5)

$$P_k\left|{}_k=(1-K_{k}H)P_k\left|{}_{k-1}{(1-K_{k}H)}^T+R_k\right.\right.$$

(6)

Compared with time domain analysis, the computational complexity of frequency domain analysis is significantly increased. Before frequency domain analysis, a fast Fourier transform (FFT) of the time domain signal is required. FFT is based on the Fourier transform theory, and the classical Fourier transform (FT) is defined as Formula (7):

$$F(\omega )={\displaystyle {\int }_{-\infty }^{+\infty }f{(t)}^{-j\omega t}dt}$$

(7)

In addition to time domain analysis and frequency domain analysis, time-frequency domain analysis is often used by researchers, which is mainly based on wavelet transform. Standard Fourier analysis can only represent the frequency characteristics of the signal very well, but can hardly obtain the time domain characteristics of the data [18]. The wavelet analysis method is a time-frequency domain analysis method formed based on the windowing Fourier analysis method. This feature analysis method can perfectly combine the time-frequency domain characteristics and has good recognition performance, but it is time-consuming to process data. The recognition performance of combining time domain and frequency domain features is also good.

With the increase in the number of system features, the recognition speed of the recognition system is greatly reduced, this is because some features of the recognition system are redundant. In order to solve this problem, researchers generally use the following methods to reduce the feature dimension: principal component analysis (PCA), linear decision analysis, and boosting algorithm. The boosting method is a method used to improve the accuracy of weak classification algorithms by constructing a family of prediction functions and then combining them into a single prediction function in a certain way. They can be used to improve the recognition rate of other weak classification algorithms, that is, other weak classification algorithms are placed in a boosting framework as base classification algorithms, and different training sample subsets can be obtained through the boosting framework’s operation on the training sample set, which can be used to train and generate a base classifier. Every sample set is obtained, and the base classification algorithm is used to generate a base classifier on the sample set so that n base classifiers can be generated after the given number of training rounds n. The boosting framework algorithm then weights and fuses these n-base classifiers to produce a final result classifier. Among these n-base classifiers, the recognition rate of each individual classifier may not be very high, but the result of their combination has a high recognition rate, which improves the recognition rate of the weak classification algorithm [19]. FFT is a fast algorithm of discrete Fourier transform, which can transform a signal into a frequency domain. It is difficult to see the characteristics of some signals in the time domain, but if they are transformed into the frequency domain, it is easy to see the characteristics. This is why many signal analyses use FFT transform. In addition, FFT can extract the spectrum of a signal, which is often used in spectrum analysis. The sampled digital signal can be converted into FFT. The FFT results of n sampling points can be obtained after FFT. The original signal and the signal after FFT are shown in Figure 3.

3.3. Overall Design of Alarm Module

The attitude angle was used as the judgment basis for detecting falls, and the attitude angle calculation formula was used. In this design calculation, whether pitch and ROL1 exceed the maximum threshold generated in the daily life of the elderly was judged respectively. If it exceeds the maximum threshold generated in the normal life of the elderly, the stage of prediction and alarm will be entered, and then the data will be continuously collected for 1 s to re-judge whether pitch and ROL1 exceed the maximum threshold generated in the daily life of the elderly. If it still exceeds the threshold, the formal alarm stage will be entered; otherwise, the warning stage will be canceled [20]. As shown in Formulas (8)–(10):

$$\mathrm{pitch}=\arctan (ax/\sqrt{a{y}^2+a{z}^2})$$

(8)

$$\mathrm{roll}=\arctan (ay/\sqrt{a{x}^2+a{z}^2})$$

(9)

$$\mathrm{yaw}=\arctan (ay/\sqrt{a{x}^2+a{y}^2}/az)$$

(10)

Pitch, roll, and yaw represent the three Euler angles of the object coordinate system (human body in this article) and ground coordinate system, respectively, in which pitch is the pitch angle, yaw is the offset angle, and roll is the roll angle.

In order to reduce the error of judgment as much as possible and prevent the unnecessary alarm caused by the error of judgment, the function of allowing the elderly to cancel the alarm is designed for the elderly to use. The specific operation is as follows: when the iPhone 6 detects the fall of the elderly through continuous data collection, it will first make a sound to attract the attention of the elderly or people around the elderly and display the “cancel alarm” button on the screen, and start to time 30 s at the same time. If the elderly presses the “cancel alert” button within 30 s, the app will also cancel sending messages to the elderly’s children, and the iPhone 6 will return to the detection state. Otherwise, the alarm will continue to sound and send alarm information to the corresponding children of the elderly.

The effective features often extracted in acceleration-based human motion pattern recognition are generally divided into time domain features and frequency domain features. Generally, the recognition rate of the combination of time domain and frequency domain features is quite high, which has been reported in many pieces of literature, but it is still necessary to extract time domain and frequency domain features. At the same time, necessary filtering and windowing processing were carried out. A sliding window with a time domain length of 128 sampling points was selected, that is, a time length of 2.56 s, 50% overlap, and a sampling frequency of 50 Hz. FFT was conducted in each window to obtain frequency domain signals [21]. In this way, the frequency we can distinguish is 0.39 Hz, and the frequency domain data of most human movements will not be lower than this value.

It is assumed that the time domain signal of a certain window of an axis is obtained, and feature extraction is carried out in time domain and frequency domain space, a series of features can subsequently be obtained.

Based on the original FFT coefficient feature, a new feature extraction method, the maximum spectral FFT coefficient feature (MFFT), is proposed. Considering that the frequency component with a large amplitude of spectrum reflects the most important component of motion waveform, it is proposed to only take the largest several spectral values as FFT features, namely the maximum spectral FFT coefficient (MFFT). The MFFT only takes M coefficients with the largest spectral values while maintaining the relative sequence of the spectrum on the frequency axis.

Figure 4 shows the comparison of the first four dimensions of the traditional FFT coefficient and the maximum spectral FFT coefficient (MFFT) proposed in the text. The FFT4 coefficient is characterized by the first four FFT coefficient values obtained in the order of increasing frequency, while the MFFT4 coefficient is characterized by the first 4-dimensional coefficient values obtained in the order of frequency values from largest to smallest while maintaining the original frequency magnitude relationship [22]. The FFRT4 coefficient is shown in Figure 4. The MFFT4 coefficient is shown in Figure 5.

According to the standard of one, the nursing grade of the disabled elderly is divided into three grades, a total of ten services. These services include home services, daycare, family care, home nursing, home and community rehabilitation services, purchase (rental) of assistive devices and improvement of home accessibility, nutritional catering services for the elderly, transportation, and long-term care institutions. There are different services or financial support for different grades. See

Table 2. Lists of disability grades and service items.

	A Mild Disability	Moderate Disability	Severe Disability
ADLs	1 to 2 items Only IADL is disabled and lives alone	3 to 4 items	More than 5 items (including)

.

In the elderly population base, the individual difference is big. According to the different levels of disability and the different needs of the disabled elderly, the corresponding care content should be different, such as simple physical medical care, psychological-emotional comfort, etc.

The smart elderly care system includes a three-step process. First of all, we need to establish a basic elderly care service information system. Based on the information system, we can apply Internet technology, sensor technology, remote monitoring technology, etc., to the elderly care life. Through the information system, we can efficiently manage and obtain information in real-time and notify and distribute the obtained information to the designated personnel in a timely manner. Secondly, through the smart elderly care service system, the elderly care information can be interconnected, and the elderly care institutions, communities, society, and other service resources can be integrated, such as health management, housekeeping services, medical services, rehabilitation care services, catering services, cultural and entertainment services, elderly care products, and travel and leisure service providers, so as to provide professional and efficient elderly care services through one entrance and by coordinating various elderly care service resources to provide the elderly with the services they need. Finally, combined with intelligent wearable devices, health detection devices, behavior monitoring systems, and other intelligent terminal devices, even AI (artificial intelligence) is applied to intelligent elderly care, and the collected elderly information is uploaded to the intelligent elderly care information system. Through calculation and analysis, the data results are sent to the children of the elderly, the elderly community, care, and medical service institutions. In case of abnormalities, early warning is given in time and corresponding services are provided.

4. Results and Analysis

The intelligent elderly care service platform provides the elderly with correct, scientific, and professional rehabilitation care knowledge and training guidance. Through big data technology, personalized rehabilitation training plans are formulated for the elderly in accordance with their own physical conditions, and systematic behavior guidance is provided for the elderly in accordance with the rehabilitation training plan [23]. First, this includes explaining rehabilitation training movement skills and matters for attention, and then carrying out a demonstration of movement essentials to ensure that the elderly learn training skills and giving guidance to assist the elderly rehabilitation training. If the elderly are bedridden or have poor physical movement ability, the aid of artificial intelligence technology can be introduced to help the elderly better adapt to rehabilitation training. At the end of each program, the training effect of the elderly can be analyzed through data perspective, and the following training plan can be reasonably arranged according to the rehabilitation of the elderly. Thus, inappropriate rehabilitation training behaviors of the elderly due to a lack of relevant knowledge and information reserve can be reduced, and the negative effects of inappropriate training on the body can be reduced, as shown in Figure 6.

The motivation factor is another indispensable condition for the elderly to actively acquire information and knowledge related to rehabilitation care. The motivation factors mainly include the user’s attitude towards preventive behavior and the related social orientation. Among them, attitude mainly refers to the user’s cognition of the preventive behavior, the expectation of the cost and benefit of the preventive behavior, and the content of the preventive behavior. Social orientation mainly includes social values and the support of people around [24,25]. The elderly with a positive attitude have the motivation to carry out rehabilitation nursing and will take the initiative to learn the correct professional way. For the elderly with a more negative attitude and in need of rehabilitation nursing, examples of positive effects of other rehabilitation training for the elderly can be shown to them. In this way, they can be motivated to learn and master scientific behavioral skills to produce good preventive behaviors, as shown in Figure 7.

Similarly, motivation can also directly affect the prevention behaviors of the elderly, such as the hospital where the elderly are eager for rehabilitation and their expectations for the future, which are conducive to the physiological and psychological rehabilitation of the elderly, and thus produce behaviors that have a positive effect on their own physical rehabilitation, as shown in Figure 8.

The ultimate purpose of the elderly’s choice of long-term care mode and the use of long-term care service is to maintain or improve the quality of life of the elderly, and the elderly will form different evaluations of their own quality of life because of their choice of long-term care mode and service use behavior. According to this evaluation, the elderly will adjust their long-term care decision-making behaviors and change their utilization degree and efficiency of pension resources for individuals, families, and society, so as to improve personal life satisfaction and pension quality [26]. See Figure 9.

After the original data collected through the network are processed through the standardization process, SPSS 22.0 software is first used to reduce the dimensionality of the standardized data. Then, according to the principle that the characteristic value of the common factor is greater than 1, when the number of extracted features reaches 7, the cumulative contribution rate is 89.412%, indicating that the extracted features reflect 89.412% of the original indicator information and can fully reflect the long-term care service supply capacity of each region as shown in

Table 3. Calculation results of characteristic value, principal component contribution rate, and cumulative contribution rate.

The Principal Components	The Eigenvalue	Contribution Rate (%)	Cumulative Contribution Rate (%)
1	16.019	47.112	47.117
2	6.44	18.471	66.041
3	2.334	6.744	72.018
4	1.917	5.664	78.584
5	1.374	4.033	88.499
6	1.361	3.774	86.317
7	1.022	3.077	89.412

.

In 2019, the CLASS survey showed that nearly 90% of the elderly choose their own home or their children’s home for retirement, while the proportion of rural elderly choosing family members for retirement reached 97%. The CLASS survey was officially carried out in 28 provinces (cities and autonomous regions) in China except for Hong Kong, Taiwan, Macao, Hainan, Xinjiang, and Tibet. The method of stratified and multi-stage probability sampling was adopted in the survey. First, county-level regions (including counties, county-level cities, and districts) were selected as the primary sampling units; second, the village/neighborhood committee was selected as the sub-sampling unit; third, the drawing sampling method was adopted in each village/neighborhood committee to sample households, and one elderly person was interviewed in each household; finally, the sample of the CLASS project included 134 counties, districts, 462 villages and households, with a total of 11,511 people. However, while institutional pension is an important supplement to family pension, its acceptance by the public is not optimistic. Among them, the overall proportion of elderly people choosing to live in nursing homes in the future is less than 5%, and the attitude of rural elderly people toward institutional pension mode is less accepting, only 1.44% of them approve. It can be seen that under the influence of Confucianism and filial piety culture, 80% of the elderly still prefer family care, while their recognition of the role of social support is low. The choice of long-term care mode for the urban and rural elderly population is shown in

Table 4. Selection of long-term care methods for urban and rural elderly population.

Long-Term Care Options	As a Whole	Cities	Rural Areas
Home of one’s own	69.90%	72.65%	66.89%
Children’s home	24.20%	18.73%	30.03%
A daycare station or nursing home in the community	0.40%	0.61%	0.08%
The nursing home	3.70%	5.88%	1.44%
Others	1.8O%	2.13%	1.57%
Total	100.00%	100.00%	100.00%

.

Considering the trend of family size miniaturization and centralization, a cross-analysis was made on the variables related to the situation of children of the elderly and their pension intention. After excluding the missing value of the main variable of “willingness to choose long-term care mode”, an effective sample of 2488 elderly people was obtained. The crossover analysis of long-term care mode selection for the elderly is shown in

Table 5. Cross-analysis of long-term care mode selection for the elderly.

Information about Children	Long-Term Care Model Selection			Sample Size
	Family Endowment	Community Endowment	Institutional Endowment
Full sample	83.48%	1.13%	15.39%	2488
No children	57.14%	7.14%	35.71%	14
Only daughter	73.78%	2.64%	23.58%	492
Only son	80.89%	1.53%	17.58%	654
With both son and daughter	87.98%	0.44%	11.58%	1328

.

5. Conclusions

With the acceleration of the aging process in China, the pressure on the supply of elderly care services is increasing. To solve the problem of elderly care supply, we should not only focus on the daily care of disabled elderly, but also pay attention to the fragmented and diverse needs of elderly care. Today, with the increasingly far-reaching impact of information technology on all walks of life, the active introduction of information technology in the field of elderly care services is not only a key measure to effectively solve the mismatch of elderly care services, but also a favorable exploration of the important role of information technology in the field of public service.

In order to effectively match the supply and demand of long-term care services for the elderly in China, the government needs to give full effort to its role of guidance, regulation, regulation, and supervision in top-level design. In this process, the upper guidance should adhere to the new concept of filial piety and build a new model of old-age cooperation among government, market, social organizations, and families so that the disabled elderly can obtain comprehensive and efficient long-term care services. At the same time, it will lead China’s long-term care service system to the path of sustainable development through continuous optimization and upgrading of policy, economic and industrial development, and service supply.