Study quality.

The detailed quality assessment of the seven reviews having met the initial inclusion criteria are summarized in the S4 Table. The quality rating presented in Table 3 indicates that the four included reviews had a high level of evidence and did not have any critical weaknesses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Synthesis of results.

https://doi.org/10.1371/journal.pone.0262889.t003

Physical activity.

The review by Vancampfort et al. [25] investigated physical activity in clinical and non-clinical populations comparing “active” to “inactive” as defined by the studies as well as meeting physical activity guidelines compared to not meeting these guidelines, defined as 150 minutes per week of at least moderate or 75 min per week of vigorous physical activity. Three of the studies (two from South Korea, one from Australia) focused specifically on older people; one a longitudinal study and the other two cross-sectional.

Multifaceted behavioral interventions.

The review by Okolie et al. [22] encompassed multifaceted interventions in three settings: a) primary care for depression screening and management, b) clinical-based (pharmacotherapy and psychotherapy) and c) community-based. The majority of the interventions addressed risk predictors such as depression and in which manner the interventions varied. They included primary care-based depression screening and management programs in which training was provided to primary care physicians on assessment and management of both depression and suicidal behavior; pharmacotherapy with antidepressants; psychotherapy in the form of Problem Solving Therapy, Problem Adaptation or Supportive Therapy; telephone counseling for vulnerable older adults via a 24-hour Friendship Line or geriatric outreach via schedule telephone calls; and community-based programs incorporating education, gatekeeper training, depression screening, mental health/health education workshops; participating in social, voluntary and recreational activities; and exercising together.

Suicide.

The 2009 review on antidepressant use included SSRIs and other second generation drugs was not able to assess the association with suicide as no suicide deaths were reported in the approximately 233 RCTs. In contrast the 2015 review on antidepressant use focused exclusively on long term use (two to 11 years) of SSRIs found no statistically significant decreased risk for suicide (RR = 1.06, 95% CI 0.68–1.66) based on two observational studies (n = 100,343).

Suicide attempts, self-harm, suicide ideation.

The experimental studies reviewed by O’Connor et al. revealed a significantly lower OR with antidepressants compared to placebo for both suicide attempt and serious self-harm (OR = 0.06, 95% CI 0.01–0.58) and suicide ideation (OR = 0.39, 95% CI 0.18–0.78). In the studies included by KoKoAung et al. concerning long-term SSRIs compared to no treatment, experimental studies found no lower odds of suicide attempts (OR = 1.00, 95% CI 0.14–7.10; n = 592) or suicide ideation (OR = 0.52, 95% CI 0.14–1.94; n = 1 821). In addition, in the latter review observational studies demonstrated that those with long-term SSRIs had an increased risk of suicide attempts compared to those without treatment (RR = 1.18, 95% CI 1.10–1.27).

For its part, the review on physical activity was based mainly on cross-sectional studies comparing “active” versus “inactive” persons (as defined by each study) and those who met the recommended international physical activity guidelines (150 min per week of at least moderate or 75 min per week of vigorous intensity physical activity) versus those who did not. Older “active” persons had significantly lower odds of suicide ideation compared to “inactive” ones in two out of the three studies (67%; n = 50,745).

The Okolie et al. review on multifaceted interventions found that of the three types of interventions studied, only primary-based collaborative care was of sufficient methodological quality. This care refers to case managers and physicians working together on identifying and managing depression and risk of suicide, demonstrating a reduction of self-harm (attempts and ideation) and suicide ideation at four, six and eight months. Results at 24 months were mixed with one cluster RCT reporting no significant difference (p = 0.12) and two other cluster RCTs showing significantly lower levels of ideation (p = 0.01; OR 0.80, 95% CI 0.68–0.94).

This systematic review of reviews synthesized data from existing reviews targeting the prevention or reduction of suicidal behavior in older people. The four included reviews were assessed as being of high methodological quality, but not all included studies were sufficiently powered to study suicide or did not include this outcome due to its low incidence. This may be due to several reasons, such as the low base rate of suicidal behavior in older (2–4 attempts for each suicide death) compared to younger ages (200 attempts for each suicide in some adolescent and young adult samples)[27]; the frailty of older adults resulting in high dropout rates in studies and the difficulty in inclusion of older adults at risk for suicide in outcomes research due to their isolation in society, comorbidities and functional impairments [28]. Assessing multiple reviews enabled us to integrate data from a large number of studies and the small degree of overlap (one study) limited the risk of duplication of conclusions while ensuring the reviews were not overly selective.

Although antidepressant medication for treatment of depression in older people has been investigated over a long period of time in both experimental and observational study designs [13, 28], our review suggests that data from the 2015 review of SSRIs showed no evidence of a protective effect for suicide or suicidal ideation [24]. As SSRIs may also be prescribed for mood disorders other than depression [29, 30], there may be an indication bias. One other challenge when comparing studies on antidepressants is that the population included in RCT studies may include a population at higher risk. Thus, a matter of heterogeneity may be present. Also, the high dropout rate in the studies included as noted by KoKoAung et al. may have led to worsening of depression and therefore predispose older persons to increased risk of suicidal behavior [24]. Results from these types of studies examining outcomes for pharmacological treatment versus no treatment are difficult to interpret due to potential confounding by indication. Physicians may have been more likely to prescribe antidepressants for older persons with more severe depression, who in turn were also more likely to thus have suicidal behaviors. As the studies were observational there was no control or experimental group to control for this. As the sale of antidepressants is rising globally [31, 32], it is necessary to continue scrutinizing the data in order to optimize the use of antidepressants. Furthermore, other drugs used in mood disorders have shown some efficacy for prevention of suicidal behavior in adults, and thus warrant further study in older populations [33].

Investigating protective factors, the Vancampfort et al. review showed that physical activity compared to being inactive had protective effects on suicidal ideation in two out of three studies on older people. As this result was based mainly on cross-sectional data, additional prospective observational studies and controlled trials are needed to confirm the findings. Furthermore, as the majority of studies did not include a comprehensive physical activity assessment, data are lacking on what type and dose (i.e. length, frequency, duration) of physical activity would be optimal for older people. This information is crucial to guide prevention programs as it is not known to what extent older adults with serious physical illness and functional limitations (both associated with suicidal behavior [5]), may be physically capable of 75 minutes of rigorous exercise per week. Also, those with depression may not feel able to exercise regularly due to lack of energy and motivation. Nonetheless, intervention strategies for older people may target the importance of physical activity for prevention of suicidal ideation based on these preliminary findings. It is key that public health research focuses on the impact of general health promotion to optimize successful aging across the health continuum [34] and capture whether such an approach may synergistically reduce suicidal behavior [14].

The review encompassing multifaceted interventions aimed at addressing risk predictors for suicidal behavior and ideation found that collaborative management in primary care was associated with reduced self-harm (composite measure of suicide attempt and suicide ideation) and for reduced suicidal ideation [22]. This finding can be explained in that older people may often have physical and mental health symptoms or conditions that need to be addressed, which are optimally addressed and treated in a collaborative, holistic team approach. It is important to recognize that that the inclusion of self-harm with and without suicidal intent as well as suicidal ideation may be problematic as although they have overlapping behaviors, they display some distinct risk profiles [35]. In addition, some older adult victims may not disclose any level of ideation [36] and ideators may never transition to suicidal behavior [37]. Yet, a history of non-suicidal self-harm and of suicidal ideation were shown by meta-analysis to confer a later risk of suicidal thoughts and behavior ranging from ideation to attempts and death [38]. Thus, further research focused on improving understanding of the determination of intent and interpretation of suicidal behavior in this age group would be informative. Furthermore, no systematic attempt was made in this review to identify possible synergies between interventions, although there is evidence that combinations of preventive interventions at several levels may have potential synergistic effects, e.g. training of general practitioners to address both depression and social isolation in older patients [39]. Previous research highlights the provision of general practitioner services due to reaching out to primary care prior to suicidal behavior and that services are accessible and integrated within other care elements to be effective in management and prevention [40].

In all four of the reviews included, the sustainability of the interventions and evidence of long-term effects beyond 24 months is unknown at this time. Longitudinal, large scale studies are required to assess this, and we identified only one multicenter study that demonstrated a reduced risk of repeated self-harm for persons of all ages combined receiving a psychosocial intervention in the short (one to five years) and long-term (10 to 20 years) compared to those receiving no intervention, as well as a reduced risk of death by suicide in the long-term; but the results were mixed when focusing on older people 50 years of age and older [41]. As the reviews identified in this study were few in number, it was not possible to stratify by gender.

Directions for future research.

This systematic review of reviews indicates there is a need for additional large-scale studies in order to build the evidence base of “what works” in suicide prevention among older people in order to intervene and alter the suicidal trajectories of older people [43]. This includes addressing the social and political determinants of suicidal behavior for a whole-of-society approach [44] given the widening gap in inequalities in suicide with higher rates among disadvantaged social groups and communities [45]. Our review identified no review addressing inequalities in late life suicide. This highlights the lack of a ‘fair opportunity of mental well-being,’ part of the World Health Organization Mental Health Action plan 2013–2020. To date, little data were found on interventions targeting high risk groups such as older men and older people who do not contact service providers, indicating the need for further investigation as to how to effectively reach these groups. Studies are also needed to examine the effect for the oldest old (≥75 years), as well as the effect of universal, population-based suicide interventions, e.g. suicide-specific funding for interventions tailored to older people. A number of interventions with promising results in other age groups or areas of mental health require testing in older people, such as cognitive behavioral therapy [46] or telemental health applications [47]. Furthermore, due to the recent meta-analysis finding of a significant effect size for multilevel interventions (different healthcare settings or domains and by different providers) on suicide and suicide attempts in persons of all ages, large scale studies are needed to determine if this is also of added value for older persons [48]. Lastly, as older people are an increasing population, the concomitant increase in life expectancy related to this demographic change means that aging-related neurological illnesses will continue to increase, such that older people may be at increased risk of suicidal behaviour. Interventions are needed to assess risk of suicide in older people newly diagnosed patients with neurological diseases [49].