Our study is the first to include mental health service delivery and personal factors such as demographic, clinical, and health-related factors to predict HRQoL through mental health service use among older persons with DD. HRQoL is an essential long-term outcome of service use. In addition, we examined predictors of HRQoL while controlling for subject correlation of patient outcomes within the same hospital.
First, this study reveals that the HRQoL of older persons diagnosed with DD who received service in a general hospital in the last six months to one year tended to be high. It might be inferred that older persons with DD had a state of health at a moderate level due to the utility index being in the possible range of -0.283 to 1, 81.37% with no problem in the self-care dimension, followed by the usual activities dimension (58.25%). In addition, the EQ-5D index was almost higher in all dimensions among older persons with continuous mental health services use than those with non-continuous service use. The lowest three dimensions of the perceived state of health include pain/discomfort, mobility, and anxiety/depression. It was supported by a report that older persons diagnosed with DD exhibited lower HRQoL concerning WHOQOL-BREF dimensions of physical health, psychological, social relationships, and global QoL [28]. In addition, [29] examined the EQ-5D index in eight patient groups of chronic diseases in six European countries (mean age = 51.9 years); participants with depression perceived no problem, most in the self-care dimension.
Predictors of HRQoL among older persons with DD
The findings reveal that mental health service use did not mediate organizational profiles and HRQol among older persons with DD. HRQol of those with DD was associated and affected by personal factors, including sex, age, comorbidity, ATDS, social support, and sufficient income, by both analyses with EQ-5D-5L and EQ-VAS. The finding of adjusted coefficient multivariable GEE analysis confirmed this finding. It is possible to explain that the HRQoL is a multidimensional and multilevel concept [31] covering various dimensions of life. The HRQoL in the current study comprised five dimensions: mobility, self-care, usual activity, pain/comfort, anxiety/depression, and evaluations of life. Mental health service use in the general hospital only provides drug treatment and psychosocial counseling related to depressive problems. As a result, mental health service use is not a mediator to the HRQoL. The results reveal that policy and the health service systems must improve intervention or expand the service to meet the long-term goal of HRQoL of older persons with DD.
According to organization groups, the high-resource organization has independently affected the EQ-5D-5L significantly; after adjusting for other personal factors, it was not affected significantly. Similarly, using EQ-VAS as the dependent variable, the higher resource organization was not affected by EQ-VAS significantly. The resource of an organization, such as nurse-patient ratio, nurse competency, and appointment reminders, was recommended for improvement.
Regarding personal factors, younger age, male sex, greater number of comorbidities, sufficient income, ATDS, and perceived social support were all statistical predictors of better EQ-5D index (p < .05). Only younger age, sufficient income, and ATDS affected health status measured with the EQ-VAS (p < .05). This finding is consistent with EQ-5D-5L utility and EQ-VAS scores reported for other countries indicating the fundamental different measurement of HRQoL [32]. On the other hand, the EQ-VAS is an adjectival rating scale that uses verbal descriptions of the respondents' health. The older persons had difficulty comparing scores between 100 (best imaginable health) and 0 (worst imaginable health). The result also accords with a recent systematic review and meta-regression analysis showing that test-retest reliability in EQ-VAS was inconsistent between those ≤ 50 years and ≥ 50 years [33].
Perceived sufficient income, ATDS, and social support had a higher HRQoL, measured by the EQ-5D-5L and EQ-VAS while controlling for subject correlation patient outcome within the same hospital in the multivariable model. This result is an explanation by the economy or income, which is essential for living in daily life, for spending to meet essential needs, or for transportation when persons have a visit for mental health service use at the hospital. This economic factor was consistent with the findings of a previous study that monthly income was significantly associated with EQ-5D-5L and EQ-VAS among older Chinese persons with depressive symptoms [2]. Regarding ATDS, Gabriel and Violato [34] concluded that knowledge of ATDS might significantly impact adherence to antidepressants. Also, a study of depressed patients who received service in a psychiatric hospital in Thailand revealed that knowledge of depression, the trustworthiness of antidepressant treatment, continued medication, and appointments were significantly associated with QoL (WHOQOL-BREF-THAI) [35].
Social support refers to information leading individuals to believe that they are cared for and loved, have self-esteem and are valued, belong to a community, and have mutual obligations [36]. The support can be emotional, informational, and instrumental for convenience [26]. In addition, social support also significantly protects the QoL regarding sensory ability, autonomy, ability, and intimacy of older persons [37]. Therefore, social support could promote the wellness and health of older persons with DD. Primarily, social support mediates the role of psychological resilience and its effect on the QoL among older people [38]. Consistent with previous studies, Kong, Zhang [38] found that social support and psychological resilience were positively associated with physical and mental HRQoL after adjusting for gender, education level, marital status, household registration, and chronic diseases among older migrant adults in China.
Similarly, previous studies on older persons with depression in the community reinforced that social support improved HRQol [37, 39]. This study found that older persons with younger age had a higher HRQoL than those with older age. This association may be because depression tends to increase in older persons due to developmental deterioration of physical and psychological aging and pathophysiological neurotransmitter disorders, especially in older persons [40], and changes in psychosocial factors such as stresses of life [41]. This finding is consistent with previous studies that poorer HRQoL by EQ-5D-5L and EQ-VAS were associated with higher age [9].
Older persons with more comorbidities had less HRQoL than those with less comorbidity, measured by the EQ-5D-5L except for EQ-VAS. A possible explanation for this result might be related to physical illness. Seniors' comorbidities may result from degenerative pathological deterioration in physical abilities that could impact their complicated illness, treatment, and care, making their health increasingly compromised. As a result of having comorbidities, older adults with depression feel more suffering from physical pain. Consistent with a previous study by[9], among older people with depression in one-year follow-up in Norway, those who had a higher number of comorbidities had lower HRQoL in EQ-5D-5L and EQ-VAS than those with less comorbidity.
The marital status of older persons was associated with EQ-5D-5L, except EQ-VAS in univariable analysis controlling for within-subject correlation between patient outcomes within the same hospital; however, after coefficient adjustment, the marital status was not associated with EQ-5D-5L and EQ-VAS. Namely, older persons who were married had a greater EQ-5D-5L than those with another marital status, thereby dependent on the influence of organization group, age, sex, sufficiency income, ATDS, social support, and severity of depression. In comparison with a previous study, which agreed with this result [42], researchers found that the association of marital status with quality of life (SF-26, physical pain) was independent of the influence of sex and psychiatric history among older persons with depressive disorders who received mental health service in the university hospital in Thailand.
The severity of depression at the first visit was not significantly associated with HRQoL after 6–12 months when using adjusted coefficient multivariable analysis measured by the EQ-5D-5L and EQ-VAS. This severity finding is in line with the findings of the HRQoL one year after hospitalization for treatment of depression in older persons who had experienced remission and had their HRQoL equal to the reference group of older persons without depression when adjusted for differences in socio-demographics and health conditions [9]. Typically, previous studies reported that the severity of depression at initial treatment is associated with poor HRQol in adults and older persons [2, 11, 32]. However, the HRQoL improved at six and twelve-month follow-ups and decreased with increasing depression severity. Besides, a cross-sectional study in older Thai persons with DD who received mental health services at the medical hospital for more than five years (57.8%) had HRQoL (WHOQOL-BREF-THAI) associated with the severity of depression, and it was dependent on the influence of education and perceived treatment outcomes [43]. Additionally, in a study of trajectories of quality of life in old age through a longitudinal approach, the finding showed that lower education and higher levels of depression and anxiety at baseline were associated with worse quality of life later; 42.1% of the variation of CASP-12 across age [44]. However, future studies must assess the quality of life between older persons who continued mental health service use from baseline and performed at the time of follow-up.
Based on these results, it can be concluded that older, male sex, lower age, less comorbidity, higher ATDS, having social support, and receiving services in the same hospital had higher EQ-5D-5L than those without these characteristics. This study demonstrated the mechanism of organization groups that was a variable influence on individual characteristics. This principle underpins the Behavioral Model of health service uses (BM), improving in the fifth phase [45] that applies organizational groups as a contextual characteristic. This study emphasized only organization groups and ATDs that promote the use of mental health services. The organization groups encompass the number and distribution of health facilities and staff to provide the population services such as nurse-patient ratio, nursing competency, appointment reminders, and hospital level. The advantage of organization groups in this study was considered distinct from the subtype pattern (high-resource and low-resource organization) and shared characteristics in all hospitals. This organization group could be an independent variable affecting mental health service use. Therefore, as discussed above, this study demonstrated the mechanism and evidence of the Behavior model of health services for investigating factors predicting mental health and quality of life among older persons with DD.
Strengths and limitations of the study
Although this study has several strengths, we have discovered that the organization groups with different available resources can directly impact mental health service use among older persons with DD. In addition, mental health service use is inadequate in predicting HRQoL measured by EQ-5D-5L and EQ-VAS. Only personal factors such as demographics, ATDS, and perceived social support affected HRQoL directly. The use of WLSMV in SEM revealed strong insights into the complex interplay between organization groups and personal factors influencing HRQoL through mental health service use, providing a valuable framework for future clinical research in this area.
However, the limitations of this study need to be addressed in four ways. First, the time to assess the HRQoL of the participants from six months to one year was too broad a range. Therefore, the score of HRQoL cannot identify the exact time and may vary. Second, research should follow up on the score of depressive symptoms when obtaining HRQoL to identify whether the good HRQoL was retrieved from older persons in remission. Third, this study did not assess information that can impact HRQoL, such as drug therapy and psychosocial intervention. Meta-analyses revealed that pharmacotherapy and psychotherapy effectively enhance QoL by improving depressive symptoms but produce mixed results [46]. Lastly, ideal PA model fit indices were observed (e.g., CFI = 1.00, TLI = 1.00, RMSEA = 0.00), which, while indicating a perfect fit, should be interpreted with caution due to the possibility of overfitting or the model's recent identification.