Selective prevention necessarily implies an initial step of risk factor evaluation, such that persons at heightened risk can be identified. Thus, important potential barriers to selective prevention include (1) lack of provider awareness of key risk factors for late-life depression, (2) need for tools for assessment of risk indicators, and (3) inadequate evidence base regarding which specific preventive interventions are likely to be successful among people with specific risk factors. In addition, finding ways to integrate knowledge of risk factor assessment and selective prevention approaches into routine primary care practice for older adults is another important challenge. Nevertheless, recent literature has advanced the field by providing considerable guidance regarding key risk factors. For example, Schoevers et al. [3] examined numerous selective prevention risk indicators; they included demographic factors (gender, age, living status (alone vs. with others), spousal loss) and health/medical (disability, presence of medical illness, trouble sleeping, cognitive impairment (measured via Mini-Mental State Exam)). The authors reported presence of physical or functional limitation/disability, spousal loss, and medical illness as substantial potential contributors to total risk of late-life depression; incorporating knowledge of living status and female gender improved computation of risk, particularly when all of the above were also considered in the context of presence/absence of subsyndromal depression. Moreover, selective prevention studies have an important “numbers” advantage—similar to that of indicated prevention trials: the relatively high incidence of depression among persons with key risk markers enables investigator to test interventions with strong statistical power, even with somewhat modest sample sizes. This fact was illustrated by Schoevers and colleagues [3], in which the authors were able to account for nearly 50 % of total risk of late-life depression with consideration of only a handful of factors. Indeed, research, largely generated by groups in the Netherlands and the USA, has identified that selective prevention may be one of the most efficient approaches to late-life depression prevention, as they have estimated that targeting persons at high risk for depression—based on risk markers such as medical comorbidity, low social support, or physical/functional disability—can yield theoretical numbers needed to treat (NNTs) of approximately 5–7 in primary care settings [4–7].

Incorporating knowledge of potential late-life depression risk factors, as described above, authors of several recent studies have successfully implemented selective prevention in clinical trials. As shown in Table 8.2, such trials have been undertaken in diverse samples of older persons, in diverse settings, and with a broad range of interventions (note: only trials of ≥3 months of treatment and/or follow-up are shown).

Numerous selective prevention randomized trials focused on older adults with major medical illness or comorbidity. For example, Rovner et al. [8] conducted a trial among 206 older patients at an eye care institution; importantly, all of the patients had already been diagnosed with AMD (age-related macular degeneration) and were newly diagnosed with disease in the second eye—thus, patients were at heightened threat of substantial visual loss. Patients were randomly assigned to receive either problem-solving therapy (PST) or usual care. At 2 months of follow-up, participants in the PST-assigned group had a lower odds of developing a depressive disorder than control/usual care participants (odds ratio [OR] = 0.39, 95 % confidence interval [CI] = 0.17–0.92, p = 0.03). After a total of 6 months of follow-up, however, most of the earlier observed benefits had diminished, though PST-treated participants were less likely to suffer from persistent depression (χ ² _(1,3) = 8.46; p = 0.04); thus, the authors suggested that “booster” sessions of PST may be required to yield more sustained preventive effects. In another successful example of selective prevention among medically ill older persons, Robinson et al. [9] conducted a selective prevention RCT among depression-free elderly patients recovering from recent stroke and used a three-group design: escitalopram, PST, or placebo control. Those receiving escitalopram had about one-quarter the incidence of major depression at 1-year follow-up, compared to those receiving placebo; patients who received PST had an intermediate outcome not significantly different from the placebo group in the intent-to-treat analyses [9]. However, by 6 months after discontinuation of the preventive intervention, the treatment group that had previously received escitalopram had greater levels of depressive symptoms than either of the other study groups [10]; thus, it is clear that further research is needed to identify the optimal means for achieving more enduring benefits of selective preventive interventions. In another RCT of older adults with a recent stroke or transient ischemic attack, Almeida et al. [11] reported that patients given folic acid (2 mg/day), vitamin B₆ (25 mg/day), and vitamin B₁₂ (0.5 mg/day) had a significantly lower risk of developing major depression than patients on placebo. In yet another study involving persons with cardiovascular disease, Salminen and colleagues [12] conducted a trial of health advocacy and counseling vs. usual care among 222 patients with coronary heart disease, aged 65 years and above, in Finland. The intervention consisted of a program of health advocacy, counseling, and activation intended to increase participants’ knowledge about CHD, social activities, role participation, support, and exercise. The investigators found that depressive symptoms tended to decrease in intervention group but to increase in control group among those men with moderate or above depressive symptoms at baseline; however, no similar changes were found among the women in the study.

Several of the studies summarized in Table 8.2 all focused on older participants with another major risk factor: physical impairments (in this case, due to musculoskeletal problems). Burns et al. [13] described findings from a selective prevention trial among 172 adults aged ≥60 years who had recently undergone surgery for hip fracture. Participants received either usual care or the psychological intervention (an evidence-based psychiatric nurse-led care intervention with elements including compliance enhancement, assistance with referrals to other professionals, advising on other services (e.g., social care), education about depression, and problem-solving). The intervention group participants, however, were not significantly less likely to have incident depression than those in the control group (OR = 0.40, 95 % CI = 0.12–1.30, p = 0.15), although the effect was noticeably in the direction of benefit. In a similar trial, Shyu and co-workers [14] treated 299 older adults aged 60 years and above, with hip fracture, with a comprehensive interdisciplinary care intervention vs. simple interdisciplinary care. Participants assigned to comprehensive care (including interdisciplinary care, nutrition consultation support, depression care management, and fall prevention) were less likely to be at risk of depression (OR = 0.48, 95 % CI = 0.28–0.83; p < 0.01) compared to those who received simple interdisciplinary care. In addition, Wang et al. [15] conducted a smaller-scale trial using the more novel intervention of tai chi (vs. attention control) among 40 older adults with knee osteoarthritis (mean age = 65 years). The investigators found that after 12 weeks of follow-up, participants in the tai chi group had significantly lower mean CES-D scores, compared to the attention control group (mean difference = −6.70 [95 % CI −11.63, −1.77; p = 0.009]). Finally, Phillips [16] conducted a trial among 202 older African-American adults (aged 67–75 years) with arthritis and chronic pain; the active experimental intervention was an education/pain management program. At 2 years of follow-up, those participants in the intervention group had significantly lower symptoms of depression on the CES-D than those in the control group (p < 0.01). In contrast with the abovementioned studies, which largely supported benefits of selective preventive interventions among older persons with medical illness, Rybarczyk et al. [17] did not observe significant differences in depression outcomes: in a study of 243 older patients with chronic illness, experimental intervention (mind/body wellness course, cognitive-behavioral treatment) yielded no significant differences in CES-D scores at 1-year follow-up compared to control, although the intervention appeared effective at reducing sleep difficulties.

Other selective prevention trials in late-life depression have focused on social or demographic risk factors as opposed to health factors. Two recent studies based in Japan focused on socially isolated older adults. Saito et al. [18] conducted a trial of a social isolation preventive intervention program among adults aged ≥65 years who recently relocated to suburban Tokyo (n = 63); although the intervention was effective at reducing loneliness as measured by a loneliness scale, there were no significant associations found between the intervention and depression, measured using the GDS. Similarly, Ukawa and colleagues [19] examined benefits of a Functioning Improvement Tool (FIT) home visit program among 252 home-dwelling older persons living in rural Japan and receiving preventive services or participating in a long-term care prevention project; again, there was no significant difference in mean depression scores of the intervention and control groups at follow-up.

A number of selective prevention trials have aimed interventions at older people living in institutional settings. Justine et al. [20] conducted a study using a multifaceted exercise intervention among 43 older adults, aged ≥60 years, who were residents of a shelter home (i.e., an institutional setting) in Malaysia. After 12 weeks of the exercise program, the intervention group reported improved quality of life when compared with the control group, but not lower levels of depression as measured by the CES-D. In another study based in an institutional setting, Konnert et al. [21] administered a CBT program among 64 older nursing home residents; evaluation tools included a diagnostic evaluation using the SCID (clinical depression was an exclusion criterion at baseline) as well as the GDS and the CES-D. Both groups were assessed on measures of depression before treatment, after treatment, and at 3- and 6-month follow-up points. Compared with the treatment as usual (TAU) control group, nursing home residents receiving the intervention (a CBT coping with stress intervention) showed improvement on the GDS over the 6-month follow-up on the GDS; for example, average scores on the GDS went from 14.0 to 9.4 points in the CBT group over the course of treatment and follow-up vs. from 13.4 to 12.3 for the TAU group over the same period. However, interestingly, results on the CES-D scale at 3 months were not statistically significant. Overall, however, the results from the Konnert and colleagues study suggested that brief, group-based CBT intervention can reduce depressive symptoms among nursing home residents at risk for depression. Haight et al. [22] also reported benefits of a selective preventive intervention in the nursing home setting. The investigators conducted an intervention of a life-review course among 256 newly relocated nursing home residents aged ≥60 years; the outcome was the BDI. After 6 weeks of receiving the intervention, the experimental group was significantly less likely to have developed clinical depression than the control group (p < 0.05), and a significant difference in the depression scores of the two groups was also observed after 1 year of follow-up. Similarly, DeChamps and colleagues [23] implemented an Individualized Cognition-Action program for late-life depression prevention among 49 adults aged ≥65 years and residing in long-term care; the outcome variable was the GDS-15. The authors found a significant difference in GDS scores between the intervention and control groups after 12 weeks, with the intervention group having a lower mean score (effect size = 0.37).

An increasingly popular treatment agent in late-life depression selective prevention trials are vitamin and nutritional supplements, and several major trials of such agents are summarized in Table 8.2 as well. For example, as noted above, Almeida et al. [11] reported strong benefits for depression prevention in the high-risk group of stroke survivors. Interestingly, however, in a large sample of 2,000 cardiovascular disease survivors, Andreeva and colleagues, in an investigation from the SU.FOL.OM3 trial [24] involving adults aged 45–80 years (mean age >60 years) receiving daily supplements of folate and vitamins B₆ and B₁₂ for 5 years, did not observe any association between supplementation and depressive symptoms among an apparently higher-risk group of older persons. Similarly, in another large-scale RCT, Giltay et al. [25] treated 4,116 hospital patients aged 60–80 years who had experienced a myocardial infarction (MI) with 400 mg EPA + DHA/day, 2 g ALA/day, both, or placebo; the outcome was the GDS-15. After the 40-month study period, no significant differences were found between any of the three groups and placebo in levels of depressive symptoms. Yet, in a smaller-scale (n = 50) study using nutritional supplements among older persons at risk for depression due to cognitive impairment, benefits were observed using an EPA-rich supplement (1.67 g EPA + 0.16 g DHA/day) or DHA-rich supplement (1.55 g DHA + 0.40 g EPA/day): after 6 months of follow-up, the mean GDS-15 scores in the EPA and DHA active treatment groups significantly decreased compared to those scores in the control group (p = 0.04 and p = 0.01, respectively) [26]. Therefore, some inconsistencies remain in the findings regarding observed late-life depression outcomes in selective prevention studies using vitamin/nutrient interventions, and further research—perhaps incorporating biomarkers, genetics, or other means of classification—may shed light on how these agents work and for whom.