During the Vietnamese transition to a modern society, the need for community-based long-term elderly care is acknowledged by the elderly, their households and representatives from the village and commune levels. This unanimity is the most interesting finding of this study. Willingness to use care services is affected by elderly ADL dependence, socioeconomic status, living arrangement and required payment level. The study respondents were willing to pay for care services at certain levels but indicated that society or associations have a responsibility to provide elderly care. There currently are no comparative figures on the need for community-centric long-term elderly care in Vietnam.
Older people in Vietnam prefer to receive care at home. This is in accordance with tradition and consistent with patterns in other countries . The majority of households and elderly only expect to stay intermittently in a nursing centre, such as 1-3 weeks/month, rather than using more long-term care as do elderly in many developed countries. The frequency of using services from any model was not dependent on the payment levels. This may be because willingness to pay was asked as a monthly payment rather than by episode of service use.
In spite of household representative expectations that their elderly would use services more often than the elderly themselves, the consensus is surprising. The findings indicate a trend of expansion of elderly care from family caregivers to a social network. This trend is likely the result of demographic pressures and socioeconomic transitions within the country. Willingness to use free services was 2-3 times higher than willingness to pay full price. This suggests a large gap between household needs and affordability of care for the rural elderly. Health insurance and medical fee exemption in Vietnam cover 33.5% of older people aged 60-89 and 37.1% of those aged 90 years and above . With this coverage, additional social and health policies to promote the formulation and use of community-centric care models are necessary. Further expanding the support for accessing free-of-charge health insurance for elderly people at younger ages, and including long-term care services in reimbursement schemes of health insurance would compensate for the limited affordability of care.
The elderly expect to receive professional care, including curative, preventive and rehabilitative services from a mobile team more than they expect informal care with tasks of daily living at home. "I suggest to the authorities that there be regular check-ups for old people" was a common view from the elderly. This opinion was supported by village leaders: "The best thing we can do is organize regular check-ups for the elderly". Medical doctors, medical doctor assistants, and nurses, working together with informal caregivers, will be essential for future provision of home care. Current estimates are 5.9 physicians and 5.6 nurses per 10,000 inhabitants in the public and private health sectors . Therefore, capacity building of a network between health professionals and informal caregivers should be addressed in strategies to expand community-centric elderly care. Social associations are also willing to contribute: "If they need builders, our members can help. Moreover, we can contribute some money".
The 2008 Vietnam Household Living Standard Survey  estimated that household per capita health care expenditure in rural areas accounts for 7% of the total household per capita expenditure. In the current study, elderly are willing to pay 2-4% of the monthly per capita household expenditure and their families were willing to pay 3-6% for services from each care model. Therefore, the maximum amount that respondents were willing to pay was almost equal to the per capita expenditure for health care.
Higher dependency in instrumental ADLs is related to higher WTU for a mobile team (except at full cost) or nursing centre (unless it is a free service). Dependent people may think they cannot afford the service at full cost price, and free-of-charge services may encourage the expectation of using services from both groups, regardless of need. This suggests that additional supports are needed for rural people to access enough care through future interventions.
A higher intellectual ADL index is associated with a lower WTU from all models of care, except free services from a mobile team. This may be because support for intellectual ADLs are mainly provided by family caregivers , rather than using day care, a nursing centre, or paying for services from a mobile team. People who are more dependent in basic ADLs are less likely to use services from a day care centre at any level of payment. This may be explained by the fact that physical exercises, health consultations and relaxation were examples of the services that could be provided in a day care centre rather than nursing care. Or, people in need of personal care preferred or relied on family support, particularly from their sons, daughters or grandchildren rather than from outsiders.
Being younger elderly is associated with interest in using a day care centre regardless of levels of payment. This could be because younger elderly are healthier and eager to participate in health promotion activities. The tendency among men to use more services can be attributed to rural women being more active in the care of grandchildren and housework, while men expect to care for themselves under the patrilineal and patrilocal culture  that remains strong in rural Vietnam. The observed trend of using more free services from a nursing centre or services with lower costs from a day care centre among people with low levels of education may be influenced by the lower incomes that are typical of this group .
It is notable that only 55.8% of married elderly are living with their spouse. The rest may live with the families of their children. People who live with a spouse are less likely to use services; widowed, divorced or single elders are more likely to use services. This may be explained by the spouse being one of the most important sources of emotional and practical support . Married people have a tendency of using more services that are free-of-charge or with lower costs. This may be affected by the 44.5% of married people who do not live with a spouse. People who live in the lowlands usually have more geographic and economic access to health services than those living in mountainous areas. This may lead to a lower expected use of mobile teams and nursing centres.
The need to spend time generating income as well as a better general health status among people still working at older ages  may explain their lower expectation of using day care centre services. People who live under the national poverty line or belong to the lower wealth quintiles are more likely to request free services and less likely to use services with payment requirements. This indicates a need to subsidize access to care among the rural elderly, especially those in poor households. Poverty status is only related to use of services with payments, but household wealth is related to use of both free and paid services. This suggests that the household short-term economic status has less effect than long-term status on willingness to use care models. Whether the poorest quintile is less likely to use free services from a day care centre than the middle quintile, or is more likely than the poorer and middle quintiles to use services with full costs from a mobile team and day care centre is unknown.
"Three main limitations are lack of budget, knowledge and guidelines" was pointed out by a representative at the commune level. This implies that more must be learned about the implementation process. The current study identified a number of activities that would improve care models. However, information is limited about how to implement programs with constrained resources. Therefore, a well-designed pilot intervention is needed that focuses on the development of an intervention, the implementation process, and attitudes of users and providers.
Some methodological issues should be considered when interpreting and discussing this study. First, community-centric models of elderly care, especially mobile teams, are rare in Vietnam. Respondents had little or no experience in using or paying for such services. Therefore, pilot interventions are needed. Second, the analysis of willingness to use and pay for services is limited by certain socioeconomic determinants and does not currently cover the health issues that face the elderly. Third, ADLs do not cover all disability domains. Therefore, this assessment of elderly care needs may underestimate other care needs for functional impairments that were not assessed. Fourth, the economic data collected by the repeat census in 2007 may fluctuate by season in rural areas. Fifth, the study respondents were not provided with estimates of the prices. Most may have no experience in paying for these kinds of services. Thus, they might interpret levels of partial or full prices differently. Sixth, this cross-sectional survey could not detect causal relationships between willingness to use the service, ADL indices and socioeconomic factors.