Mechanisms of brief contact interventions in clinical populations: a systematic review

All studies included in our original meta-analysis [5] were examined to establish the main mechanism proposed by study authors. Based on this review, the first author (AM) then proposed several overarching main mechanisms that encapsulated BCIs in these studies. These were reviewed by the co-authors, who each assessed the extent to which the mechanism proposed accurately reflected those in BCI studies. Thus, the process used to identify main mechanisms was iterative. Assessment of BCI mechanism was examined independently from the evaluation of the intervention itself. A range of terms were used to refer to suicidal behaviour and self-harm in BCI papers. We did not alter or change the original terms used by study authors.

The provision of social support was the most commonly reported mechanism (Table 1). The protective mechanisms thought to underpin social connectedness relate to perceived care and support from others, which is thought to improve coping responses to stress [37].

The idea that BCIs improved social connections or provide an additional source of social support was suggested in nine articles. This was expressed in a variety of ways. For example, Motto and Bostrom [3] suggested “being joined to something meaningful outside oneself as a stabilizing force in emotional life”, while Carter and colleagues [32] suggested that a postcard intervention may provide a “gesture of caring to counteract feelings of loneliness”. Vijayakumar and colleagues [36] described their telephone intervention as “enhancing a feeling of ‘connectedness’ and a feeling of being ‘cared for’”. While the effectiveness of the BCI was assessed independently from proposed mechanism, it was noted that all three of these papers reported that receipt of BCIs was associated with a reduction in suicidal behaviours.

Several papers also suggested that the BCIs provide a way for participants to communicate their distress and made them feel as though they are worth listening to. For example, Cotgrove and colleagues [24] described their intervention as “provid[ing] a way to obtain help…to communicate distress, [and] indicates that individuals are interested in their wellbeing”, while Fleishmann and colleagues [13] suggested that “systematic follow-up contacts gave the patient a feeling of being seen and heard by someone”. Postcard studies also typically invited participants to contact clinicians, which may have contributed to the feeling that staff were open to hearing their concerns.

BCIs may also lead to improved suicide prevention literacy, which we define as an individual’s knowledge about suicidal behaviours or self-harm (e.g., risk and protective factors), what help is available and/or where to find this information, and how to access this help at times of crisis. Increased suicide prevention literacy has been used to explain results of trial papers that have reported increased contact with health services for self-harm in the intervention group (albeit being non-significantly associated in some trials—see Table 1) [21, 27, 35].

The possibility that BCIs may work by increasing suicide prevention literacy was reported in eight articles (Table 1). Kapur and colleagues [27], found an increase in admissions to hospital in the intervention group compared to controls, suggesting that the BCI intervention may have potentially led to a “reduced threshold for help-seeking or improved engagement with services engendered by receipt of the intervention”. Bennewith and colleagues [21] reported that receipt of BCIs lead to a greater willingness to contact general practitioners for help in subsequent crises. Cotgrove and colleagues [24] found that none of those who used the Green Card to readmit themselves to hospital engaged in a repeat episode of self-harm, suggesting that these participants may have benefitted from the ability to access clinical services during a time of crisis afforded by the Green Card.

Studies finding reduced self-harm have also highlighted suicide prevention literacy as a possible mechanism of BCIs. A postcard study set in Tehran [26] found significant associations between an individual’s belief that the postcards were ‘helpful in the prevention of suicide’ and the proportion of self-reported suicidal ideation during the 12 month follow-up period. Specifically, those reporting no belief in the helpfulness of the postcards were almost twice as likely to report experiencing ideas of suicide compared to those with at least some belief in the helpfulness of the cards. There were also significant associations between belief in the helpfulness of these cards and suicide reattempts in this trial. An emergency card intervention also suggested that these cards can help participants improve their knowledge about where, and when, to seek help [28]. This implies that not only can BCIs inform participants about services available to them, including the possibility of hospital re-admission at times of crisis, but also reminds participants that they are welcome to contact these services if required.

Process evaluation from a number of other studies suggested that participants found BCIs to provide useful information and were associated with an increased likelihood of help seeking [4, 29, 35]. For example, Hassanzadeh et al. [35] reported that their BCI intervention increased participants’ recognition of the need to ascertain support from various sources including outpatient/inpatient services, relatives, and friends or by telephone contact. Robinson and colleagues [4] found that participants liked receiving BCIs and that they did make use of the health-promotion strategies recommended. However, no significant effect of the postcard intervention was found on attempted suicide, although participants in both the intervention and control groups improved on measures of mental health over the course of the study.

Participants may have also learnt more effective coping behaviours, which was reported in one trial [31, 32]. The authors of this postcard trial inferred that some alternative behaviours may have been learnt as they documented a reduction in self-poisoning readmissions and psychiatric hospitalisations over a sustained 5-year post-intervention follow-up period. Suggested examples of these alternate behaviours included improvement in coping strategies, emotion regulation, impulse control, and self-understanding.

Observational studies suggest that those with established social relationships are less likely to engage in self-harm and suicide than those who are socially isolated [38]. For example, a study on people hospitalised following a suicide attempt reported that those with greater family and peer connectedness were less likely to have re-attempted suicide after 12-months than those with less connectedness [39]. At a population level, evidence from two national samples (in the US and UK) suggested that perceived support from friends and family was associated with decreased likelihood of a lifetime suicide attempt, controlling for a variety of related predictors [40]. Studies with self-harm patients have emphasised the importance of social support post-discharge [41, 42]. The proposed protective effect of social support aligns with evidence from studies in mental health, which have reported the importance of social support in protecting against depression [37, 43] and psychosis [44]. Intervention studies have also found that peer-delivered social support “befriending” interventions have beneficial effects on depressive symptoms [37].

There are a number of ways in which social support may be protective against suicide. First, social support may have a stabilising effect to buffer against stress [3]. Hence, persons with greater social support may perceive less need for mental health services, which suggests that supportive relationships might be a substitute for or a complement to formal treatment [45]. Other researchers suggest that those with more frequent social support contact are more likely to access mental health services following a stressful life event than those with infrequent contact (i.e., social support expedites access to services), and are less likely to require specialist psychiatric services (i.e., suggesting social support plays a stress reduction role) [46].

Further, and as noted above, BCIs provide a way for participants to communicate their distress and conveys the sense that others are interested in their wellbeing. The opportunity to talk about factors that contributed to the self-harm episode is described as a important aspect of treatment in past qualitative studies [42], which supports this as being an important part of the BCI intervention. Therefore while social support is a likely mechanism, researchers in this area acknowledge that there is a need for more work into the causal mechanisms that explain why and how social support may be protective in BCI studies.

A recent cross-sectional survey among adolescents who seriously considered suicide indicated that those who sought help had greater knowledge about the availability of help resources and an understanding of these could help with their problems [47]. This aligns with studies from public health which have argued that improvements in mental health literacy lead to greater awareness and knowledge about where to seek help, and greater willingness to seek help [48, 49]. It is worth noting that previous research has suggested that suicidal populations have similar levels of knowledge about mental health problems and their treatment as people in the general community [50].

BCI interventions may also change perceptions of the care process, particularly because they seek to provide support months after a person’s initial presentation to a hospital. For example, studies have shown that those who have sought help for self-harm generally have a negative opinion of hospital-based services [51] and perceive these services as unhelpful due to poor continuity of care [52]. However, other studies have shown that adolescents who engage in self-harm and were offered therapy following discharge from hospital rated it positively [53]. Thus, the offer of help following treatment for self-harm (such as a BCI) may lead to the person viewing treatment in general as more supportive. Alongside this, BCIs may be effective in modifying attitudinal barriers and stigma, making it more likely for people to contact services should they need them. These explanations align with the process evaluations discussed above, which reported that participants liked receiving BCIs and found the messages they provided helpful.

One group of authors speculate that BCIs may involve the learning of alternative behaviours [31, 32] and cite past research using dialectical behaviour therapy (DBT), which has been associated with a reduction in suicide attempts, lower rates of hospitalisations and ED visits, and greater adherence to treatment than cognitive behaviour therapy delivered by experts [54]. However, presently it is unclear if alternative behaviours are learnt during BCI interventions and, additionally, whether these behaviours are connected to the BCIs themselves, as few trials have measured mechanisms of change for these interventions. As we mention above, examples of possible alternate coping behaviours may include: improvement in coping strategies, emotion regulation, impulse control and self-understanding. Some of these (e.g., self-understanding or improvement in coping strategies) may go hand-in-hand with greater suicide prevention literacy, which suggests that these mechanisms could overlap or, at the very least, be mutually reinforcing.

There are certain caveats and limitations that need to be taken into consideration in interpreting the findings from this review. First, there may be other BCI mechanisms not covered here. To a large extent, insights of this review are reflective of the degree to which intervention researchers articulated mechanisms for BCIs. It is also necessary to acknowledge that proposed mechanisms may operate differently for different subgroups within these BCI trials. For example, several studies have documented a variation in effectiveness between those with a history of multiple episodes of self-harm and first-time attempters [20, 21, 25, 30‐33]. Other studies report that BCIs may benefit women but not men for some outcomes [26, 32, 35]. We acknowledge that there may also be variations based on whether individuals had expressed intent to die (e.g., self harm versus suicidal behaviours with intent) and depending on whether they had previously been in therapy. Related to this, the proposed mechanisms may operate differently depending on the specific type of therapeutic model (e.g., DBT versus psychodynamic therapy) a person has received. However, as BCI studies have generally not been powered to look at these subgroup differences, it is difficult to assess how meaningful these effects may be in clinical practice. There may also be differences between written versus verbal forms of BCIs. For example, our original meta-analysis [5] suggested that postcard interventions may be particularly effective as a form of BCI. Last, we would acknowledge that there have been additional reviews published since the time of our original publication [5]. We did not include this because it was not in our original review.

Other limitations of this review relates to its scope and strict inclusion criteria, which meant only a relatively small number of studies were eligible for assessment. We acknowledge that the strict inclusion criteria limited the number of studies we included in this review. Last, the mechanisms that underpin letters and postcards may be different from that underpinning Green Cards. In saying this, we did find that there were similarities in the proposed mechanism across BCI studies.

BCI, crief contact interventions; ED, emergency-department; SH, self harm