2. Materials and Methods
2.1. Data Source
Since there are few contents involving LTC in the research database such as China Health and Retirement Longitudinal Study (CHARLS) [37
], we conducted our own case research in four regions—namely, Chengdu, Chongqing, Guizhou, and Hubei province—which are the main representative regions in the middle and upper reaches of the Yangtze river.
We adopted telephone survey, WeChat and QQ group survey to investigate the needs of LTC among the elderly in Chengdu, Chongqing, Guizhou, and Hubei province. The survey, conducted by the Chongqing Technology and Business University from March 2018 to December 2018, consisted of about 15 questions (please note that some variables were not included in the empirical analysis due to too many missing values) that took interviewees about 10–20 minutes to complete, and 1787 samples were recovered, of which 1308 were valid and over with a response rate of 73.1% excluding missing values. Online or verbal consent was sought from interviewees before the survey and no need for ethical approval. All investigations are conducted anonymously to respect privacy.
The sampling process consists of three steps as follows: (a) setting up a sampling framework for each region’s administrative district; (b) any district was subdivided into census districts and listing all the elderly residents in each affected district; and (c) the investigators recruited the elderly in each block using a random sample method.
Multiple logistic regression was selected to examine the relationship between underlying factors and LTC service needs of the elderly. The latest version of Anderson’s model was incorporated into the logistics model by adding predisposing characteristics, enabling factors, need factors, and psychosocial factors. In the control of the first three groups of variables, the hierarchical model was adopted to analyze the susceptibility, initiative, and demand of each factor. Besides, the needs for LTC was stratified by gender, age, and education with multiple logistic regression analysis to further explore the differences of influencing factors among different groups (male and female, young and old).
2.2. Measurement Method
At the beginning of the questionnaire, we first briefly explained the basic concept of LTC to the interviewees (long term care is continuous care over a long period of time for people with chronic illness, such as cognitive impairment, or impairment, known as functional impairment), and then start asking questions.
Our paper measured LTC needs in older adults with a simple question: "which LTC way do you want to choose?" Alternative answers were: 1 (home care), 2 (community-based care), and 3 (institutional care). Based on the Anderson model (the latest version), the study evaluated four independent variables sets: predisposing characteristics, enabling factors, need factors, and psychosocial factors. Many implementation processes and methods refer to the previous research literature, such as Fu et al. [38
] and Xu et al. [37
2.2.1. Predisposing Characteristics
These included age (Below/above age 69, question: How old are you?), gender (female/male, question: What’s your gender?), education level (Below Bachelor’s degree/Bachelor’s degree or above, question: What is your highest learning experience?), as well as marital status (married or not, question: What is your marital status?). In order to master the differences between different areas, the study also set "regions" (1 = Chongqing, 2 = Guizhou, 3 = Hubei, 4 = Chengdu). Unlike Fu. et al. [38
] divided education level into primary school or below, junior high school or senior high school, and College or above, we divided it into below Bachelor’s degree and Bachelor’s degree or above considering the higher and fast improved education in China in the last 40 years.
2.2.2. Enabling Factors
It is included: income level (question: Which of the following is your annual income?), quantity of children (question: How many children do you have?), and frequency of connection with children (question: How often do you contact the child?). Individual income or quantity of children was measured by continuous variable method.
2.2.3. Need Factors
The need factor consists of two variables: IADL (Instrumental Activity of Daily Living) and quantity of chronic diseases. some sub-items of IADL were evaluated by daily life activity scale. The quantity of illnesses was calculated by question as follows: "how many chronic diseases do you have?" Higher scores represented interviewees who had more illnesses especially chronic diseases.
2.2.4. Psychosocial Factors
Psychosocial factors include variables as follows: intergenerational ties, unmet needs for LTC, and self-image evaluation. (a) Intergenerational ties were measured by the following question: "How are you getting along with your children?"(answer: 1 = very poor, to, 5 = very good). The higher the score, the closer the intergenerational relationship. (b) Unmet needs for LTC were evaluated from four aspects: living surroundings, medical treatment and spiritual life. The question” Do you need the following care services” is multiple choice (0 = none,1 = only one, 2 = two, 3 = All of them). (c) The assessment of aelf-image referred to Bai et al. [39
], who came up with the Chinese version of the Self-Image of Aging Scale. A higher score meant a more positive self-image.
This study used the latest version of Anderson’s model to explore the LTC needs of the elderly in China. We added psychosocial factors to predisposing characteristics, enabling factors, and need factors in the old version. From the sample selected from the main representative regions—namely, Chengdu, Chongqing, Guizhou, and Hubei province—in the middle and upper reaches of the Yangtze river, there were some important findings.
Firstly, predisposing characteristics play a very important role. Many factors were analyzed, for example, we support the effect of marital status as many other studies [40
]. Marital status is an important factor, the preference for institutional care among the elderly who are currently unmarried is particularly strong (2.4801), which is very consistent with the reality. The discovery of regional differences may be affected by the diversity in the geographical situation, economic development, and cultural background of the four sampled regions. For example, compared with Chengdu, Chongqing has the highest preference for institutional care (1.3030), while Guizhou has the highest preference for home care (1.2782). These may be related to the level of economic development in different regions and the degree of market opening to the outside world.
Secondly, in terms of enabling factors, the research results are consistent with other research results [41
]. The elderly people who are close to their children are more likely to “age in place” regardless of whether they receive family care or community-based care. As shown in Table 3
, when them contact infrequency with children, the proportional relationship respectively was 0.7150 (home care vs. community-based care) and 2.5756 (institutional care vs. community-based care), but when they have some contact with children, the proportional relationship will be increased to 1.4532 (home care vs. community-based care). In fact, even community-based care services need to be complemented by viable care provided by family members. Because the unity between the elderly and their children is a key factor to enable the elderly to continue to live or ‘age in place’ in a familiar living environment. Surprisingly, the effect of income on long-term care pattern is very small (1.0227; 0.9996).
Thirdly, in terms of need factors, people with more chronic diseases more prefer institutional care (1.0899), which contradicted some research results [43
]. It may be due to the restricted number of IADL and the disease in the sample, indicating that most interviewees were in good health. The elderly who choose community-based care were more likely to show unmet needs for care than those who choose home and institutional care. On the one hand, the limitation of community nursing sources limits the fulfillment of the elderly with community nursing sources and LTC needs unmet. On the other hand, the elderly who choose community-based care are more possible to show the needs of community-based care than those who choose home care.
Fourthly, the role of psychosocial factors in elderly LTC service needs was abundantly demonstrated in this paper: Following to control the other three groups of factors, the variance of 2.835% can be explained by psychosocial factors. This confirms the significant role of psychosocial factors in influencing LTC needs, consistent with previous research results [15
]. Meaningful correlations between psychosocial factors and LTC service needs indicate something. Table 3
also reveals the model changes among the four regression models. According to the demand of LTC, four kinds of multiple logistic regression models are established. The change in model fitting is calculated. Inducers have the greatest explanatory power for LTC demand differences (pseudo-r2 = 0.0861). The addition of enabler increased the interpretation capacity by 2.7% (chi-square = 98.6076). After adding the demand factor, the improvement was 1.2% (chi-square = 110.039). The joining of psychosocial factors improved by 2.7% (chi-square = 136.5735). In terms of self-image evaluation, the elderly with the higher self-image evaluation has the lower preference for community-based care, which may be related to the fact that Chinese elderly people pay more attention to ‘personal face’ (personal dignity) and do not want to show their shortcomings to others, especially those around them. Self-image evaluation is composed of several factors such as general physical health, attitude towards life or psychosocial states [39
]. People who have positive self-perceptions about their physical health, attitudes to life, and social status recognition mostly think they have the ability of taking good care of themselves, are more active to involve social interactions, and therefore willing to accept home care. In traditional Chinese cultural values, the elderly attach great importance to the views of the society or others on them. They will try their best to show better aspects to the outside world, and whether they can be taken care of by their families is one of the important aspects.
At the end of the empirical analysis, regression analysis by gender/age/educational level group was conducted in order to further examine the differences in factors affecting LTC service needs in the elderly. (a) From the perspective of gender, it is a difference in the mainly influencing factors of LTC service needs for men or women among the elderly. As shown in Table 4
, taking marital status as an example, the difference between male and female is significant when them currently not married (OR = 1.4080 and 4.1832 in institutional care vs. community-based care). The role of men or women in social construction and the gender division in labor market may help to clarify the LTC needs of elderly people affected by gender differences. (b) In terms of age, the influencing factors for the two-age group (age in 60–69 and above 70) are completely different. This may be because in China, the life expectancy is about 75, the elderly over 69 years may detect the urgency of the demand for LTC while the elderly under 69 years think this demand is more distant. (c) From the view of educational level, psychosocial factors—especially unmet care service needs and self-image evaluation—played an important role in the choice of the elderly. Unmet care service needs (institutional care vs. community-based care, 1.9558) and self-image evaluation (institutional care vs. community-based care, 1.1796), have a prominent effect on the choice of the elderly with Bachelor’s degree or above compared to the elderly with below Bachelor’s degree (1.0848; 1.5245).
The growing trend of an aging population stressed the urgency of the task in the fund collection of long-term care (LTC) services for the disabled elderly. This research, though based on findings in the Chinese context, could afford a reference value for other countries, especially those that similarly emphasize familial relationships.
This study aims to explore the factors that affect LTC needs of the elderly in the frame of the latest Anderson Model, which added psychosocial factors to predisposing characteristics, enabling factors, and need factors in the old version. In fact, few researches studying the role of psychosocial factors in influencing factors of LTC needs of the elderly and testing educational-related differences in this subject in China. The sample is selected the main representative regions—namely, Chengdu, Chongqing, Guizhou, and Hubei province—in the middle and upper reaches of the Yangtze river. It found some interesting phenomena, but the results may be slightly inaccurate due to the selection of regions and the design of questionnaires, which needs to be further improved and deepened in the future research. In addition, we did three more sub-analyses—namely, by age, by gender, and by educational level. They are useful supplement, but perhaps more factors, such as health insurance purchase status or long-term care insurance (LTCI) purchase intention, can be taken into account in future studies.