Effectiveness of a village-based intervention for depression in community-dwelling older adults: a randomised feasibility study

The suicide rates in older adults in South Korea are much higher than those in other Asian and Western countries; the rate was 47.7 per 100,000 older Koreans in 2017 [1], compared with 13.8 in the United States, 22.9 in the European Union, and 21.1 in Japan [2, 3]. It is also roughly double that of the younger population in Korea (24.3 per 100,000) [1], and is now regarded as a major social issue. In Korea, suicide rates are higher in rural areas than in urban regions [4], which might be due to the large numbers of elderly people living in rural areas. Suicide prevention strategies tailored by age and residential area are needed; however, only a few trials have examined this issue in older adults living in rural areas [5].

Among the risk factors for suicide in older people, depression is an important modifiable factor [6‐8], and has been a target of several suicide prevention trials [9‐11]. However, late-life depression is usually under-recognized and under-treated compared with depression in younger adults, because of the propensity for masked depression and comorbid physical illnesses [12‐14]. Therefore, previous interventions have focused on the surveillance of depression by primary care physicians, or have targeted older adults who already had a clinical diagnosis of depression [9‐11]. However, small rural areas usually have insufficient medical services, including a lack of general physicians (GPs) and psychiatrists. Furthermore, a meta-analysis showed that GPs were less successful in identifying depression in older adults than in younger people [14], whereas a recent review reported that community outreach services effectively reduced the risk of suicide [15].

Therefore, this study tested the feasibility of a self-sustainable community-based intervention for depression, targeting older adults living in rural villages. We developed a 12-week program that was implemented by the community mental health service (CMHS) team; in this program, community-dwelling healthy adults played the main role in preventing depression in at-risk older adults. Rural areas have some beneficial features, such as relatively small physical and psychological distances among individuals; this may facilitate socializing among neighbors. Although 243 CMHS centers have been installed throughout Korea, each service team has multiple roles, caring not only for depressed older people but also for individuals with other chronic mental disorders. Therefore, we developed a short-term, sustainable depression intervention for small rural areas that encouraged the participation of healthy residents who could continue in their depression prevention role after the program ended.

This pilot study investigated the efficacy of a community-based multilevel intervention for late-life depression in two small rural areas in Korea. Two villages were allocated randomly to the intervention group and active control; all older adults living in the two villages were included in either the intervention program or received standard CMHS care, and the effectiveness of the protocol was assessed in representative older adults according to age- and sex-stratified random sampling.

Of all older adults living in the two villages (n = 451), 160 (35.5%; intervention, n = 82; control, n = 78) were included in the assessments. Figure 2 shows a flowchart of participant enrollment and assessment; enrollment continued until the target sample size was achieved. During the study period, 17 individuals in the control group were lost to follow-up due to poor health, lack of time, or relocation to another area. No adverse events were reported during the study period.

Table 1 shows the baseline sociodemographic and clinical characteristics of the participants by group. Overall, the participants had a mean (SD) age of 74.2 (5.7) years and 5.1 (4.1) years of education; 94 (55.8%) were female. Older adults living in the intervention village were less educated and had poorer socioeconomic status, reflected by greater their higher rate of Medicaid coverage; they also had a higher rate of lifetime history of suicidality at baseline, higher baseline depressive symptom scores, and lower baseline social network scores compared with those in the control group.

The linear mixed-effects models did not show a significant group × time interaction for the primary outcome measure (Table 2). By contrast, among individuals with high depressive symptom scores at baseline (SGDS-K ≥ 6), a greater decrease was observed in the intervention group than in the controls, although the difference did not reach significance (mean difference [SD], − 2.41 [3.83] and − 0.77 [2.59], respectively; β [SE] = 3.05 [2.86], t (28) 1.066, p = 0.296).

Rates of severe depression (SGDS-K ≥ 8), incident major or minor depressive episode, and incident suicidality did not differ between the intervention and control groups at post-intervention period. However, in the logistic regression analysis dividing subjects into low and high baseline depressive symptom groups, trends toward a decreased risk of severe depression (SGDS K ≥ 8) and incident major or minor depressive episodes were observed in intervention group subjects with more depressive symptoms at baseline (Table 3).

Regarding secondary outcomes, a significant group × time interaction for scale of social network as seen in the intervention group (Table 2). Post-hoc analyses showed a greater increase in the LSNS-K in the intervention group between baseline and follow-up (mean difference [SD], 3.67 [7.32]) than in the control group (0.53 [10.61]; β [SE] = 3.08 [1.34], t (153) = 2.301, p = 0.023). No other significant differences were observed between groups.

The village-based intervention for late-life depression conducted in this study did not reduce depressive symptoms in older adults overall; however, the program tended to lower the risk of progression to severe depression among at-risk older adults, and increased the social network more than five-fold compared with standard CMHS care.

To our knowledge, this is the first multilevel depression intervention for community-dwelling older adults using representative samples. Most previous community-based depression or suicide interventions measured the effectiveness of their protocols according to health-related indices or suicide rates at the general population level [25‐29]. They were generally effective in improving mental health indices (e.g., rates of mental health service use) or lowering overall suicide rates; however, the measures deemed most important, and how the protocols affected the at-risk population specifically, were unclear. We assessed depression and suicidality in participants directly, according to demographic and clinical factors both before and after the intervention. The results suggest that our community-based protocol was effective for increasing social bonding in older adults although depressive symptoms were not reduced by the intervention: it only showed a tendency of preventing the progression to severe depression in the at-risk population.

Our 12-week intervention did not reduce depressive symptoms in the older population overall, which was probably due to the inclusion of large numbers of healthy elderly individuals. Based on growing evidence that multilevel approaches are more effective for managing depression and reducing suicide rates [11, 30‐32], we used such a design in our community-based intervention to prevent late-life depression in a healthy population and at-risk individuals. However, the primary outcome (changes in depressive symptoms) might not have been sufficient to demonstrate the benefits of our short-term program, because there was little room to for improvement of SGDS-K scores given the low average score at baseline. On the other hand, although our program was open to all elderly residents of the intervention village, the pre- and post-intervention assessments were performed only in representative samples, after excluding individuals who had difficulty participating in the interview. Therefore, it is possible that the effectiveness of our intervention was underestimated due to exclusion of the data for unhealthy older adults who underwent the intervention but were not analysed. Additionally, although the older adults in the study generally exhibited good cognitive health, even mild impairments in cognitive function could have influenced the reports of depression. This was especially true for K-CIDI-defined depressive disorders based on DSM-IV criteria, because this tool uses somewhat complex questions. Although additional adjustments, including for baseline MMSE-KC scores, did not change the primary results of the logistic regression models, the inclusion of clinical assessments of depressive disorders by psychiatrists should be considered in future studies.

Meanwhile, there was an even increased tendency for severe types of depression to be reported in the intervention group. It is possible that this type of disclosure was more common because baseline SGDS-K scores were significantly higher in the intervention group than the active control group, even though the scores were within the normal range for both groups. Additionally, older adults in the intervention group might have been more likely to report their mental health problems after interacting with the local committee members and/or the CMHS staff. In general, older Asians are unwilling to disclose their mental health problems to other people because they tend to perceive emotional problems as a sign of weakness, a lack of discipline, and/or a lack of willpower [33]. With considerations for this cultural context, we cautiously suggest that the participants in the intervention group might have been less reluctant than the active controls to report their severe depression to the researchers after adapting to expressing their emotional status during the intervention program.

In the present study, there was a tendency for the intervention to prevent the progression of severe depression in at-risk elderly individuals (i.e., those with moderately high depressive symptoms at baseline). These results suggest that a more targeted intervention for at-risk older adults would be beneficial, even though the results did not reach the level of significance. This finding is consistent with those of previous community- and clinic-based trials targeting individuals at high risk of depression. However, GPs were mainly responsible for screening patients for depression and referring them for psychiatric treatment in most such studies [9, 10, 34]. There are few medical services in many rural areas of Korea, including Yeoncheon, in which only 29 physicians and one psychiatrist cover a population of 44,187 [16]. Moreover, a primary care system involving GPs is not the standard in many non-Western countries, including Korea. Therefore, our protocol has a possibility to be particularly useful for managing high-risk older adults living in small rural communities, although further studies including more numbers of villages are needed.

Our community-based intervention for geriatric depression strengthened the interpersonal networks of elderly individuals, as measured by the LSNS-K; this scale evaluates not only the size of an individual’s social network, but also the frequency and feasibility of networking. A lack of social relationships is associated with various health problems [35, 36], including depression [37‐39] and even mortality [40, 41]. Depressed older adults tend to have more fragile social networks, which was reported to be related to a greater propensity toward suicidal thoughts [42]. A previous community intervention that used a group-based approach to increase the social networks of older adults showed promising results [43], consistent with our study.

Social support may protect older adults against worsening depression, by mitigating psychosocial stressors [37]. A large community-based observational study showed that infrequent social interactions was significantly associated with depression in older adults [39]. People typically have less contact with friends and colleagues as they age, so it is important to re-establish interpersonal relationships, especially with close friends or neighbors [43]. Our program aimed to strengthen the autonomy of village-dwellers; healthy local committee members were responsible for monitoring the at-risk older adults, and group activities were used to encourage reciprocal care. Encouraging older adults to assume active roles in the community might increase their social networks. Our promising intervention should be assessed in the context of long-term prevention of late-life depression in future studies.

This study had several limitations. First, although we randomly allocated two villages as the intervention and active control, two group sizes may restrict the conclusions. Second, while all elderly population living in two villages were targeted by intervention or usual care, respectively, only one third of them participated in the baseline and follow-up interviews even though they were randomly selected from each group as representative populations. Third, the blinding of both researchers and participants was only partial. Lastly, the intervention was conducted over a short period compared with other depression prevention programs.

Our village-based intervention for late-life depression, led by healthy community residents and CMHS staff, was effective in increasing the social networks of older adults and might prevent severe depression among at-risk elderly individuals. We believe that this protocol is potentially useful for preventing depression and suicide in older adults living in small rural communities, which usually have insufficient medical services. Future studies that include more villages, employ a long-term follow-up period, and apply more targeted strategies for at-risk older adults will be needed to validate the findings of the present feasibility study.