Page 49 - IJPS-11-5

International Journal of
Population Studies Need for social services among disabled older Chinese

ordered logistic regression models for the entire study urban differences are statistically significant.
sample. We then replicated the analysis for rural and Table 3 reports the results from multilevel ordered
urban residents separately. At Level One, the units are logistic regression models for the entire study sample
observations at each wave, nested within Level Two units as well as for urban and rural sub-samples. Among
which are respondents. Each respondent contributes predisposing factors, the results for the entire sample show
multiple observations across different waves. At the person that living in a rural area was associated with 9% and 13%
level (i.e., Level Two), only the intercept was included in higher odds of needing basic care and social connection
the analysis. services, respectively, compared to living in an urban area.
All independent variables in the multilevel logistic The results for the total sample further show that, except
regression models were included simultaneously. Missing for cases where Han ethnicity and being married were
data among all covariates were generally low (<2%); to associated with higher odds of social connection service
minimize potential bias due to missing values, we imputed need, other predisposing factors related to demographics
the mode for categorical variables and the mean for and occupation were not significantly associated with the
continuous variables. Alternative imputation approaches need for social services when enabling and need factors
were also assessed (e.g., multiple imputations) and the were simultaneously modeled. In the urban sample, being
results were nearly identical. We also assessed possible married was associated with higher odds of need for both
multicollinearity among variables and found that all types of services. In the rural sample, all predisposing
variance inflation factors were less than 3. All analyses factors were not significant.
were performed using STATA version 17.0. For enabling factors, the results were more complicated.

3. Results Economic independence is the enabling factor that only
affected urban residents and was associated with reduced
Table 1 presents the descriptive statistics of the study odds of the need for both basic care and social connection
sample: older adults with ADL disabilities in China. In services by about 29 – 30%; Some enabling factors affected
the sample, about half of the respondents resided in rural both rural and urban residents: Coresidence was associated
areas. The mean age was 94.8 for rural respondents and with 16% and 20% reduced odds of the need for basic
94.1 for urban respondents. Men accounted for 29.2% care services in rural and urban areas, respectively. The
of rural respondents and 47.8% of urban respondents. availability of social services also affected both rural and
Overall, access to basic care and social connection services urban residents. Availability of basic care services was
was very limited in both rural and urban areas, with an associated with 27% increased odds of the need for basic
average of fewer than one available service out of four. care services in the rural sample and 31% increased odds
Significant differences were observed between urban and of the need for basic care services in the urban sample, but
rural residents in terms of socioeconomic status and care it was not associated with the need for social connection
resources. For instance, urban residents generally had services. Availability of social connection services was
higher levels of education, were more likely to hold white- associated with 60% and 63% increased odds of the need
collar jobs, had better access to medication, and enjoyed for such services in rural and urban areas, respectively; it
greater economic independence. By contrast, they were was also associated with 16% and 24% increased odds of
less likely to coreside with their children or have a family the need for basic care services in rural and urban areas,
member as their primary caregiver. All observed rural– respectively.
urban differences were statistically significant. Among need factors, poorer self-rated health was
Table 2 displays the percentage of respondents who significantly associated with 15 – 16% increased odds of the
reported a need for eight types of social services. More than need for basic care and social connection services in urban
half of respondents required assistance from each type of areas only. Unmet need for care was associated with 12%
service; the need was greatest (>60%) for home visits, and 16% greater odds of the need for basic care services
psychological counseling, health education, and personal in rural areas and urban areas, respectively. Severe ADL
care. Rural older adults reported a greater need for all eight disability was associated with 13 – 17% higher odds of the
social services compared to their urban counterparts. In need for basic care and social connection services in urban
particular, rural older adults had a much higher need for areas only. With respect to control variables, compared to
home visit service (80.3% vs. 73.5%) and neighborhood the need in 2005, there was an increase in the need for both
relations (61.3% vs. 56.7%). In terms of the two summary basic care and social services from 2008 to 2018.
variables, rural residents had a higher need for both types An additional analysis comparing regression
of services compared to urban residents. All the rural– coefficients (results not shown, available upon request)

Volume 11 Issue 5 (2025) 43 https://doi.org/10.36922/ijps.448