Advancing Knowledge About Suicide Prevention Media Campaigns

In recent times, media campaigns have been gaining popularity as a means of preventing suicide (Ftanou et al., 2016). Media campaigns differ in their target audiences and their messaging, but essentially fall into two groups. Campaigns in the first group are aimed at the general population – or subgroups within it – and are designed to raise awareness about suicide and convey messages about openly discussing suicide and recognizing the signs that someone may be going through a particularly rough time. Campaigns in the second group target suicidal individuals themselves, emphasizing that they are not alone, encouraging them to talk to family and friends, and/or pointing them to sources of professional help. Our recent work in this area has prompted us to consider a number of issues relating to these two groups of campaigns.

Arguments and Assumptions Underpinning Media Campaigns

Campaigns that target the general population have received considerable support. The argument for these campaigns usually goes as follows: Suicide has been shrouded in silence for too long, and if people understood it better and felt more comfortable talking about it, then the suicide rate would be reduced (Legge, 2011). These campaigns are guided by the assumption that communities know very little about suicide. In fact, this is probably not quite true. Studies that have examined suicide literacy at a population level have tended to show that people are aware of suicide as a public health problem and can recognize many of the warning signs (Reavley & Jorm, 2011). However, these studies also suggest that people have difficulty translating this knowledge into action and are cautious about approaching those they may be concerned about (Rossetto, Jorm, & Reavley, 2014). The corollary of this is that suicide prevention campaigns that target the general population may not actually need to improve community understanding of suicide, because baseline levels are already quite high; instead they should focus on how to equip people with the skills to support someone who is in a suicidal crisis.

Campaigns that target suicidal individuals themselves are based on the notion that these individuals often feel that they are isolated and have no-one to turn to, believe that their problems are insurmountable, and do not know where to go for professional help. Again, these assumptions warrant examination. For example, there is evidence from community surveys that suicidal individuals are much more likely to use mental health services than the general population, even compared with people with mental health problems (Johnston, Pirkis, & Burgess, 2009), suggesting that many of them do know where to go for help. Even those who do not use services may not be prevented from doing so by a lack of knowledge; factors like stoicism and a preference for self-managing problems are likely to play a role here, and these in turn may be influenced by experiences of stigma (Griffiths, Crisp, Jorm, & Christensen, 2011; Reynders, Kerkhof, Molenberghs, & Van Audenhove, 2015). Again, this has implications for the content of campaigns, suggesting that addressing knowledge may not be sufficient and that beliefs and motivations may need to be considered.

Campaign Messaging

Messages for the general population are, by design, different from messages for suicidal individuals themselves. An example of the former is: "Start the conversation today," which was the tagline in a campaign run in New Zealand in 2013 that encouraged viewers to initiate discussions about suicide (Ftanou et al., 2016). An example of the latter is: "It takes the courage and strength of a warrior to ask for help … If you are in emotional crisis call 1-800-273-8255, press 1 for veterans." This is taken from a campaign run in the United States in 2010, which urged at-risk veterans to seek help (Pirkis et al., 2016). There are also examples of messages that target both groups simultaneously, such as: "Youth suicide: There is hope … If you or someone you know is suicidal, call 1-800-273-TALK," which was the centerpiece of a more general campaign conducted in the United States in 2010 (Pirkis et al., 2016).

It is not always clear how these messages have been developed, but presumably they are often the result of collaborations between suicide prevention experts and communications professionals. It would be interesting to know whether any of these messages have been formulated with recourse to other tried-and-true areas of suicide prevention. For example, school- and university-based programs that are designed to equip students to recognize warning signs for suicide and intervene with at-risk peers often incorporate mnemonics such as LAST (Listen, Act, Show support, Tell an adult), ACT (Acknowledge, Care, Tell), and QPR (Question, Persuade, Refer). These programs have been shown to yield positive outcomes, at least in the immediate term (King, Strunk, & Sorter, 2011; Schilling, Aseltine, & James, 2016; Strunk, Sorter, Ossege, & King, 2014; Tompkins & Witt, 2009), so the mnemonics might be adaptable for use in media campaigns that target the general population.

It is also worth considering the potential impact of given messages on the target audience and on other audiences. We recently ran a workshop where we asked 21 people with professional expertise in or lived experience of suicide to develop campaign messages for three different target audiences – the two described earlier (people in the general population who might be in a position to intervene with someone at risk of suicide and people at risk of suicide themselves) and people bereaved by suicide. Our workshop participants noted that messages that might have the desired effect for one audience might have unintended consequences for the other audiences – for example, those who had been bereaved by suicide might be made to feel worse by messages encouraging the general population to recognize warning signs, because they might feel that they had missed these signs themselves (Ftanou et al., 2016). Care obviously needs to be taken to ensure that campaign messaging is safe for both target and nontarget audiences.

The Effectiveness of Media Campaigns

A number of studies have been conducted that have looked at the effectiveness of suicide prevention media campaigns. These have varied in quality. At the lowest quality end, there are studies that have only collected data after the campaign, usually via survey. At the highest-quality end, there are studies that have collected data before and after the campaign and employed additional design refinements (e.g., additional data collection time points, comparison groups) and/or multiple data sources, and randomized controlled trials (RCTs) of campaign elements (e.g., community service announcements that form the core of many campaigns; Pirkis et al., 2016).

Collectively, these studies suggest that exposure to campaigns can lead to improvements in awareness and understanding of suicide and, often, attitudes toward it. However, they also suggest that behavioral outcomes (such as increases in help-seeking) are more difficult to achieve in a consistent fashion. The jury is also out on the extent to which they can have an impact on suicide rates, although some studies have demonstrated that this outcome can be achieved (Pirkis et al., 2016).

Ideally, we should identify campaigns that seem to work and campaigns that seem to have minimal effect, and examine the differences between them. At a macro level, it would be useful to see whether one of the aforementioned two groups of campaigns is more effective than the other; this would allow us to determine whether we should put our efforts into campaigns targeting the general population or campaigns targeting suicidal individuals themselves. At a more micro level, it would be desirable to examine the specific messages delivered through successful and unsuccessful campaigns within each group. Doing this could inform the message content of future campaigns. At present, this is not possible because evaluations of previous campaigns provide varying levels of detail on campaign messages (Pirkis et al., 2016).

Improving the Evaluation Effort

Media campaigns certainly show promise as an intervention to prevent suicide, but further evaluation is required before definitive statements can be made about their effectiveness. A recent systematic review of seven key suicide prevention strategies concluded that there was insufficient evidence of the benefits of media campaigns, and that the paucity of RCTs in this area was the rate-limiting factor (Zalsman et al., 2016). We would certainly agree that there is scope for further RCTs of various campaign elements, and in particular would advocate for trials that pit specific campaign messages against each other. These sorts of trials can be conducted in laboratory-like conditions, where participants are randomly allocated to receive particular messages.

We would also argue, however, that it is trickier to conduct RCTs of whole campaigns. This is because campaigns are complex interventions and the conditions in which they are run are less than ideal from an evaluation perspective. Like other universal interventions, media campaigns are, by their very nature, delivered to whole populations. Evaluators often have no control over the reach of given campaigns and are usually unable to randomly allocate subpopulations to intervention and control conditions. Sometimes there are opportunities for using quasi-­experimental designs, where the intervention subpopulation is set (e.g., if a campaign has been developed for a particular city or state) and control subpopulations (similar cities or states) can be identified. Occasionally, it is possible to stagger the roll-out of a campaign across several cities or states, using a modified stepped-wedge design. We would argue that all suicide prevention campaigns should be evaluated, and that campaign developers and program evaluators should work together to create the most rigorous evaluation designs possible.

Where it is not possible to conduct RCTs – and even where it is – evaluations can be strengthened in certain ways. For example, explicating the program logic of the campaign in question will be helpful; clarifying the causal pathway by which the campaign would be expected to achieve its desired outcomes will help to determine whether it is successful in doing so. Similarly, triangulating or using multiple methods and data sources to assess the effectiveness of given campaigns will strengthen the conclusions that can be drawn. These approaches are commonplace in other areas of program evaluation and are particularly useful in relation to complex interventions.

Conclusion

Media campaigns are gaining traction as a means of combatting suicide. It is important to distinguish between campaigns designed for the general population and campaigns designed for suicidal individuals themselves. It is crucial that we get the messaging right in each of these sorts of campaigns. Currently our knowledge about which messages work is insufficient, so it is imperative that suicide prevention experts and campaign developers collaborate closely and draw on evidence from elsewhere in the suicide prevention field. Ongoing rigorous evaluation of whole campaigns and their individual elements is vital in order to ensure that they do good, not harm.