Receiving threatening or obscene messages from a partner and mental health, self-harm and suicidality: results from the Adult Psychiatric Morbidity Survey

Violence, abuse, and bullying are widely recognised as factors contributing to both the onset and continuation of mental disorder [1, 2]. England’s Department of Health and Social Care (DHSC) recommends that healthcare professionals ask patients about current and historic exposure to violence and abuse, to identify risk factors, establish safety, and support recovery [3‐5]. A dose–response relationship has emerged with the risk of poor mental health being greater where experience of violence and abuse has been more pervasive [6]. People exposed to one type of violence and abuse are more likely to also experience other types, at the same time or subsequently [7].

One of the most common forms of violence in adulthood is that from an intimate partner (IPV). The World Health Organisation lists different types of IPV, including physical violence, sexual violence, emotional or psychological abuse, and controlling behaviours, involving isolation, monitoring, and restricted access to resources [8]. In the UK, the Serious Crime Act 2015 was the first to recognise controlling or coercive behaviour in the context of an intimate or family relationship as an offence [9], and the Domestic Abuse Act 2021 specifically recognises economic and technology-related abuse [10, 11]. However, a tension has been noted between such public policy and the gendered reality of domestic violent crime [12].

Emotional abuse is known to predict poor mental health as strongly as physical or sexual abuse [13] and includes insults, belittling, intimidation and threats, delivered face to face or communicated in other ways. The Malicious Communications Act 1988 had made it illegal in England and Wales to send or deliver letters or other articles for the purpose of causing distress or anxiety, and in 2001 the legislation was updated to also apply to electronic communications [14]. From March 2019 to March 2020, 36% of all domestic violence-related stalking and harassment offences recorded by police in England and Wales were malicious communications [15]. Police forces, however, have been found to often not record—or to mis-record—such reports [16]. Crime statistics therefore significantly underestimate the extent of abusive behaviours conducted through communications, including digital communications, in the population, with further changes to likelihood of reporting occurring in the context of COVID-19 restrictions [17].

Global developments in communication technology have the potential to enable malicious communications to become increasingly pervasive, immediate, and intrusive [18]. While the UK government’s recent Online Harms white paper acknowledged this, it made no reference to its occurrence in the context of IPV nor accounted for the growing risk from Internet-connected devices [19]. Instead attention has focussed on threatening or obscene messaging in other situations: such as hate speech [20], children’s online behaviour [21], and conduct on social media. Furthermore, research on sexting [22] and cyberbullying has tended to focus on peer-related bullying rather than IPV. Malicious communications can form an aspect of technology-facilitated abuse (‘tech abuse’), which includes the monitoring of victims through networked camera systems, stalking via software and tracking tools frequently referred to as “stalkerware” [23, 24], and other forms of “smart” technological systems such as Internet-connected household appliances to coerce, control, and harm others [25]. Limited recognition has been given to these newer forms of abuse, especially in the context of emotional abuse and coercive control from a current or previous partner. In 2020, the UK Law Commission [26] launched a consultation on reforming the Malicious Communications Act 1988 and Communications Act 2003, in light of developments in online communication and recognising the inherently gendered nature of online abuse and its potential for causing harm.

While England’s mental health survey did not ask specifically about tech abuse, it did include questions on receipt of repeated threatening and obscene messages from a current or previous partner, spanning both technology-facilitated and other forms of communication. The aims of this study were to establish:

We carried out a secondary analysis of data from the 2014 Adult Psychiatric Morbidity Survey (APMS), a general population survey of the mental health of adults resident in private households in England. A stratified random probability sampling design was used, based on the selection of addresses from the Postcode Address File (PAF). This involved multiple stages: sampling primary sampling units (PSUs); addresses within selected PSUs; and one individual aged 16 or over from each selected address. People living in communal or institutional establishments, in temporary housing, or sleeping rough, were not in scope. Fieldwork took place May 2014–September 2015, with verbal informed consent.

Seven thousand five hundred fourty-six participants were interviewed, a response rate of 57%. Weights were developed to take account of selection probabilities and non-response, in order to render results representative of the household population. Interviews averaged 1.5 h and were conducted in people’s own homes (or elsewhere, if preferred) by trained research interviewers. The questionnaire was largely administered face-to-face using computer-assisted interviewing. Some information considered more sensitive was collected by self-completion for increased privacy, with participants keying their responses directly into a laptop. Interviewers assisted respondents with literacy or eyesight difficulties. 7058 participants carried out the self-completion component: factors associated with declining the self-completion included being male and being older [27]. Further methodological details are published elsewhere [28, 29].

Our results suggest that in England around 96% (n = 6857) of adults aged 16 or over have had at least one partner or spouse. Men (5.4%, n = 122) were more likely than women (2.7%, n = 79) to have never had a partner (p < 0.001). Subsequent analyses are based on participants reporting at least one partner in their life, to avoid results being confounded by relationship history.

Repeated threatening or obscene messaging causing fear, alarm or distress is a relatively common but rarely researched form of partner abuse experienced by one person in fifteen in England. Consistent with research on cyber IPV [37], those experiencing such messaging are more likely to be young, female, and living in socioeconomically disadvantaged circumstances, although it is also experienced by men, older people, and across socioeconomic groups.

Those experiencing threatening or obscene messaging are often also exposed to other forms of abuse and violence too. This cross sectional dataset provided a snapshot in time. We could not distinguish abusive messaging that presaged other forms of IPV (for example, as a feature of dating abuse) from that which occurred alongside other forms of IPV (perhaps to control a current partner when they are physically out of reach). The results are also consistent with research highlighting abusive messaging as a continuation of coercive control post-separation, after someone has left an abusive relationship [38]. This includes the use in ‘revenge porn’ after a relationship ends. While popular discourse around the non-consensual sharing of sexual images has focussed on the motivations of a ‘jilted sexual partner’ [39‐41], this analysis refocuses attention on the harms to victims.

People who received repeated threatening or obscene messages from a current or previous partner in the past year were more likely to experience anxiety and depression, to engage in self-harm, and to be suicidal. These associations remained elevated even after adjusting for confounding for current socioeconomic disadvantage and exposure to a range of types of violence and abuse across the life course. Lindsay and colleagues [42] have shown how abusive messaging from an intimate partner increases the odds of experiencing fear as a result of victimisation, and that, in turn, fear was associated with an increased risk of depression. Our results are consistent with a trauma-informed understanding of mental health, and support the evidence base by demonstrating the additional effect that such communications can have. Health care professionals carrying out safety assessments should therefore enquire routinely and explicitly about messages and communications, and these should refer to any received from a previous as well as a current partner. Voluntary and statutory support services engaging with people who have left abusive relationships should establish whether survivors still have ongoing contact with former partners, and the nature of any communication, to ensure threats, intimidation, harassment and monitoring are not missed.

This study was unable to examine technology-enabled abuse in isolation. To respond to emerging risks—with implications for the frequency, immediacy, and visual content of threatening or obscene messaging—technology-enabled communications should be asked about in research and during safety assessments and accounted for during safety planning. New telephone numbers, contact details, and ideally devices, should be issued to those at risk [43]. Mobile providers and other suppliers should facilitate blocking other people across platforms. As technology may change the ways in which abuse is mediated, future surveys will need to adapt to ensure that the scale and impact of such experiences in adulthood and childhood is captured, including on longitudinal studies. The rapidity of such changes poses challenges for research, for service responses, and for victims.

This is the first analysis using representative English mental health survey data to present the national prevalence of threatening and obscene messaging from partners, together with its social and mental health correlates. The survey was based on a national probability sample and high-quality self-completion procedures in the context of a face to face interview. The identification of threatening or obscene messaging, however, was based on a single question, which did not distinguish between online and offline forms of communication, whether the perpetrator was a current or former partner, or whether messages were felt to be threatening or obscene. In the context of what Herman [44] calls ‘domestic terrorism’, however, sexual comments and pornographic images from a current or previous partner will often be experienced by women as both threatening and obscene. The question also asked only about the receipt of threatening or obscene messages that had actually provoked fear, distress, or alarm in the participant: in consequence, it failed to capture people who had received such messages but felt unaffected by them. The lower rate in men, for example, might have reflected the fact they were less likely to experience distress from such messages, or were less likely to disclose having felt distress. Recent (past year) experience was reported by relatively few (124) participants, and some of the analyses were accordingly underpowered. The findings, however, echo gendered dynamics identified in the context of the sending and receiving of sexualised images. Female teens receive far more unsolicited messages and images than males and are also more stigmatised for sending, for example, nude pictures in return [45]. The Conflict Tactics Scale (CTS) [34], which informed the questions used to capture other forms of partner violence and abuse, has been identified as problematic. The scale amalgamates non-equivalent types of gendered violence (for example, implying that being ‘slapped’ and ‘pinned down’ are equivalent) [46], and fails to capture context or impact [47].

The sampling frame excluded those living in residential or institutional settings or sleeping rough; and people living in domestic violence refuges and shelters were unlikely to have been selected. However, the proportion of the population not living in private households is small [48] and would have had little impact on the prevalence estimates and associations examined. While a survey response rate of 57% is in line with similar surveys, bias due to non-participation is inevitable. While cross sectional surveys using high quality general population samples are the best data source for prevalence and for profiling (understanding nuanced socioeconomic circumstances and the relationship with other types of violence and abuse), conclusions cannot be drawn about causality.

This study supports earlier research in finding pronounced associations between emotional abuse and poor mental health and suicidality. When asking about violence and abuse, clinicians, law enforcement, and service providers need to ensure specific enquiry about threats, intimidation, and harassment from partners who may not be present in the home. Such enquiry, as well as future research, must address the potential for ongoing abuse from former partners, who may use remote communications to continue to exert threats and control [37]. Abusive messaging is one strategy of control that partners, especially former partners, use when their victim is not physically available [49]. In the context of the unintended impacts of ongoing physical distancing measures to mitigate against Covid-19 transmission [50‐52], addressing abusive messaging has only gained in urgency.

Mental health services and future research should cover both online and offline modes of communication. Malicious communications in the context of IPV should feature in policy and legislation on both domestic violence and online harms. Our results support calls for technology companies to publish annual transparency reports to assess the prevalence of harmful content on their platforms. Hopefully through such means more data and evidence on online and offline facilitated forms of communication-enabled abuse will become available and will help to contextualise future research in this area.