Problematic Smartphone Use Mediates the Pathway from Suicidal Ideation to Positive Psychotic Experiences: a Large Cross-Sectional, Population-Based Study

Abstract

The present study followed the newly hypothesized “suicidal drive for psychosis” suggesting that psychosis may be consequential to suicidal ideation (SI) and behavior and attempted to explain parts of the pathway between these variables. To this end, we aimed to test whether problematic smartphone use (PSU) has an indirect mediating effect in the cross-sectional relationship between SI and positive psychotic experiences (PEs). Lebanese community young adults (N=4158; 64.4% females; mean age 21.91±3.79) have been invited to participate to a cross-sectional, web-based study in the period from June to September 2022. After adjusting for potential confounders (i.e., the living situation, marital status, household crowding index, economic pressure, cannabis use, other drugs use, and past history of mental illness other than psychosis), we found that higher levels of suicidal ideation was significantly associated with greater PSU, which was also positively and significantly associated with more positive PEs. Finally, greater suicidal ideation was significantly and directly associated with more positive PEs. Our findings suggest that SPU may be regarded as a potential target for prevention and intervention in psychosis. Clinicians, educators, and school administrators ought to give greater attention to PSU among vulnerable young people who present with SI.

The dimensional approach of the psychosis spectrum stipulates that psychosis exists on a continuum ranging from subclinical psychotic symptoms and experiences seen in non-clinical individuals from the general population to more severe and debilitating psychotic disorders (Van Os & Reininghaus, 2016). Psychotic experiences (PEs) include delusional beliefs and perceptual abnormalities and are highly prevalent in healthy youth (Fekih-Romdhane et al., 2020; McGrath et al., 2016; Yates et al., 2021), with a median lifetime prevalence estimated at 7.2% (Linscott & van Os, 2013). Although attenuated and/or infrequent, PEs may be potentially distressing and disruptive, interfering with daily activity and functioning (Carey et al., 2021; Healy et al., 2018). PEs have also been shown to contribute to elevated healthcare costs and increased use of mental health services (Bhavsar et al., 2018; Rimvall et al., 2020). PEs have been demonstrated to be clinically useful, not only in conferring increased risk of developing later psychosis, but also in predicting a wide range of psychopathology and behavioral problems (Healy et al., 2019; Knight et al., 2020). Therefore, a thorough investigation of the psychopathological factors related to PEs may deepen our knowledge of the mechanisms behind a heightened vulnerability to psychosis and offer new opportunities for prevention and intervention in psychosis.

The Relationship Between Suicidal Ideation (SI) and Positive PEs

A strong evidence has shown that SI and behavior are highly prevalent among people with diagnosed psychotic disorders (Harkavy-Friedman, 2009), and people at clinical-high risk (CHR) (Fekih-Romdhane, Abbassi, et al., 2021). Today, it is also sufficiently well-established that the greatest risk of experiencing SI and behavior occur before any contact with mental health services (Ayesa-Arriola et al., 2015; Granö et al., 2011; Welsh & Tiffin, 2014). For instance, numerous studies have reported high occurrence of SI before any clinical assessment among young people at risk for psychosis (Hui et al., 2013; Hutton et al., 2011). An Italian study found that 8.6% of CHR individuals presented with a history of suicide attempt(s) prior to clinical assessment and enrollment in an early intervention program, whereas attempted suicide occurred significantly less frequently during follow-up (Preti et al., 2009). The study also revealed that a family psychiatric history in first-/second-degree relatives was associated with a heightened risk of attempted suicide (Preti et al., 2009), suggesting that there may be a genetic vulnerability for suicidality among these CHR individuals (Tsai et al., 2011) and also that suicidal behavior could be an inherited genetic trait that precedes the onset of psychotic manifestations. Besides, other researchers have demonstrated that SI and behaviors are commonly reported by healthy young individuals from the general population experiencing PEs (Koyanagi et al., 2015; Taylor et al., 2015). However, traditionally, the vast majority of previous studies have mainly focused on the pathways leading from PEs to SI and behaviors (for review, see Yates et al., 2019). Many national (Connell et al., 2016; Fisher et al., 2013; Kelleher et al., 2014; Kelleher et al., 2013; Martin et al., 2015) and multinational (Bromet et al., 2017; Honings et al., 2016) studies have prospectively investigated subsequent risk of suicide in individuals who first reported PEs. It has been only recently that the inverted pattern was hypothesized. In a prospective study of Danish cohort (Murphy et al., 2018), authors found that self-harm and suicidal behavior have more likely preceded the onset of psychosis and that PEs prevalence significantly increased with the increase of SI/behavior recency and severity. Later, Murphy et al. used a prospective data from the Environmental Risk (E-Risk) Longitudinal Twin involving a cohort of 2232 twins and found that suicidal ideation and behavior (any suicidal attempt or self-harm) at age 12 conferred a 2.48-fold risk of occurrence of PEs at age 18 (Murphy et al., 2022). Interestingly, authors also showed that 39% of individuals who have experienced both suicidality and PEs at age 18 reported suicidal-, threat-, or death-related content of PEs (Murphy et al., 2022). According to the suicidal drive hypothesis in the relationship between psychosis and suicide that has initially been proposed by Murphy et al., psychosis represents a form of “threat responsivity” to suicide (Murphy et al., 2018). In other words, SI and behavior have been suggested as a self-directed and internally generated “threat to survival,” which can lead to psychotic manifestations as psychological responses aimed at protecting the individual against this external threat (Murphy et al., 2018). This is plausible, particularly since positive PEs are experienced as external to the self and involve threat-related content (e.g., beliefs that others tend to have harmful intentions and that personal safety is threatened (Freeman, 2007); one’s own mind or body being under the control of external forces (Frith, 2012); pejorative and threatening auditory hallucinations or voices commanding suicide (Harkavy-Friedman, 2009)). This is also in accordance to the theory that PEs are unspecific signs of general psychopathology severeness (Loch, 2019). In light of these data, we have focused in the present paper on assessing a possible factor mediating the relationship from SI to PEs. As such, problematic smartphone use seems to be a potential candidate, being especially linked to both SI and PE.

Problematic Smartphone Use (PSU) as a Mediator of the Relationship Between SI and Positive PEs

PSU behaviors involve both process (content consumption of media such as viewing movies and browsing websites) and social (social media interactions such as messaging) aspects of usage (Elhai et al., 2017). There is limited but growing evidence that multiple facets of PSU are associated with a range of psychosis-risk indicators in young healthy people, including high psychoticism (Cao & Su, 2007), schizotypal personality traits (Fekih-Romdhane et al., 2022; Massaro et al., 2022; Truzoli et al., 2016), pre-psychotic symptoms (Rizzo et al., 2015; R. F. Tzang et al., 2015), and PEs (Fekih-Romdhane, Sassi, et al., 2021; Lee et al., 2019; Ruu-Fen Tzang et al., 2015). For instance, a longitudinal Canadian study revealed that a higher tendency for digital media use was significantly linked to greater PEs among community youth (Paquin et al., 2022). Several studies have shown that problematic digital media use precedes the development and exacerbation of PEs (e.g., perceptual and thought disturbances) (Stip et al., 2016; Torous & Keshavan, 2016). In addition, certain types of self-reported behaviors on social media (e.g., venting, posting about feelings, commenting on the posts/pictures of others, and viewing profiles of non-friends on social media) have been shown to contribute to subsequent increase in paranoia among people with and without experience of psychosis (Berry et al., 2018). Mittal et al. found that persisting patterns of problematic Internet use were closely associated with increased PEs during follow-up among young adults, suggesting the Internet as “a new environmental stressor” in the psychosis etiology that “unmasks the subtle vulnerability” (Mittal et al., 2013) according to the diathesis-stress model of psychosis (Walker et al., 2008). PSU can act as a causal factor and precipitates the emergence of psychosis through a decrease of potential protective factors, such as sleep and exercise (Griffiths et al., 2022; Odgers & Jensen, 2020), or when it turns into addiction (Lin et al., 2016; Ratan et al., 2021).

On the other hand, there is increasing evidence that SI is cross-sectionally associated with problematic Internet use and PSU in adolescents and young adults (Chahine et al., 2020; Cheng et al., 2018; Dib et al., 2021; Liu et al., 2020; Marchant et al., 2017; Sampasa-Kanyinga & Lewis, 2015; Sedgwick et al., 2019; Wan Ismail et al., 2020). The relationship between PSU and suicidality has gained research attention within the context of the COVID-19 pandemic, due to the substantial increase in both digital technologies use (De’ et al., 2020) and suicide rates (John et al., 2020). Research during the COVID-19 has, for example, documented that problematic Internet use significantly correlated with SI in Spanish university students (Villanueva Silvestre et al., 2022). A systematic review by Mamun has identified increased exposure to social media as a SI risk factor in the context of COVID-19 (Mamun, 2021). Suicidal youth may resort to their smartphones for multiple reasons, including a need to disclose their negative feelings and experiences to others on social media (Bazarova et al., 2017; Park et al., 2012), online help-seeking (De Luca et al., 2020; Seward & Harris, 2016), suicide forum use (Mok et al., 2016), seeking online support (Harris et al., 2009), and seeking for information about suicide methods (Sueki, 2013). These suicide-related Internet use have been found to have substantial negative impact on individuals experiencing SI. Beyond the immediate risk of increasing SI and encouraging suicide (Liu et al., 2020; Mok et al., 2016; Sueki, 2013; Whitlock et al., 2006), smartphone usage may easily become addictive in this vulnerable population. For example, a Chinese study found that people with SI were significantly more active on social media compared to other users (Wang et al., 2019). Another Chinese study showed that college students with SI in the past year had 2.6-fold increased risk of PSU, reported more than 5 h of current daily smartphone use, and more severe smartphone addiction symptoms (i.e., inability to control craving) (Huang et al., 2021). Similarly, a Korean study revealed that college students exhibiting SI had around 2-fold increased risk of smartphone overuse than their counterparts without SI (Kim et al., 2019). Taking into consideration all these data, it seems plausible to suggest that SPU is a relevant mediator in the SI-PEs relationship.

The Present Study

Adolescence and young adulthood is a period of proneness to smartphone addiction (Tossell et al., 2015), as well as emergence of a broad range of psychopathology, including SI and psychosis (Solmi et al., 2022). To our knowledge, no previous studies have explored PSU as mediator in the suicide-psychosis relationship. Examining and identifying the influence of mediators may help elucidate the potential mechanisms underlying the relationship between suicide and psychosis, and, in turn, the development of evidence-informed prevention and intervention programs. Based on their observations and prospective findings, Murphy et al. (Murphy et al., 2022; Murphy et al., 2018) called for additional investigation of the newly hypothesized “suicidal drive for psychosis.” To respond to this call, we aimed through the current work to test whether PSU has an indirect mediating effect in the cross-sectional relationship between SI and positive PEs.

Methods

This study belongs to a large cross-cultural, binational project involving community-dwelling participants from two lower-middle-income Arab countries, Lebanon and Tunisia (Psychotic Experiences in Arabs from Lebanon and Tunisia, the PEARLS project; N=4891). This project is aiming at describing and comparing the phenotypic manifestation of PEs in these countries, as well as validate the Arabic version of the Community Assessment of Psychic Experience (CAPE-42) scale in the Arabic language. Only the Lebanese data were analyzed in the current study.

Lebanese community young adults have been invited to participate in a cross-sectional, web-based study in the period from June to September 2022. A proportionate random sample of participants from all Lebanese governorates (i.e., Mount Lebanon, Beirut, South, North, and Bekaa) has been used. Individuals willing to participate were eligible if they were aged 18–35 years and had no past personal history of psychosis or antipsychotic drugs intake. An anonymous online survey, which contained information about the study, was shared on social media platforms (WhatsApp, Facebook, and Instagram) and used to collect data from as many participants as possible. No financial compensation was given to volunteering participants. Internet protocol (IP) addresses were examined to ensure that no participant took the survey more than once.

A total of 4158 participants enrolled in this study, with a mean age of 21.91 ± 3.79 and 64.4% females. All sociodemographic and other characteristics of the participants are summarized in Table 1.

Table 1.

Sociodemographic characteristics of the participants (N=4158)

Variable	N (%)
Gender (female; n, %)	2679 (64.4%)
Governorate
Beirut	649 (15.6%)
Mount Lebanon	872 (21.0%)
North	681 (16.4%)
South	1224 (29.4%)
Bekaa	732 (17.6%)
Marital status
Single	3889 (90.9%)
Married	379 (9.1%)
Education
Secondary or less	198 (4.8%)
University	3960 (95.2%)
Housing area
Urban	2050 (49.3%)
Rural	2108 (50.7%)
Living situation
Alone	180 (4.3%)
With family	3941 (94.8%)
With friends	37 (0.9%)
Cigarettes smoking (yes; n,%)	554 (13.3%)
Alcohol drinking (yes; n,%)	674 (16.2%)
Cannabis use (yes; n,%)	79 (1.9%)
Other illegal drug use (yes; n,%)	44 (1.1%)
Presence of mental illness (yes; n,%)	1055 (25.4%)
	Mean ± SD
Age (in years)	21.91 ± 3.79
Household crowding index	1.52 ± .75
Economic pressure	6.14 ± 2.68
Positive dimension	33.03 ± 7.19
Suicidal ideation	.47 ± 1.00
Problematic smartphone use	30.22 ± 12.19

Questionnaire

Before accessing the survey, participants had to give their digital consent after ready the introductory paragraph; the latter contained information about the following: (1) the anonymity of the answers, (2) that only researchers will have access to view any data collected and (3) that this data cannot be linked to the participant, (4) participant can withdraw from the study at any time, (5) that the refusal to participate will have no penalties, (6) that the research project has been approved by an ethics committee, and (7) he/she voluntary agrees to take part in this research project. The questionnaire was divided into three sections and was presented in the native language of participants (i.e., Arabic). It required 15–20 min to complete. The first section involved socio-demographic data of participants, including gender, age, presence of a diagnosed mental illness, and substance use. We also collected information on self-perceived economic pressure level using a 10-point Likert single item scale, i.e., “How much pressure do you feel with regard to your personal financial situation in general?” (scores range from 1 to 10, with 10 referring to a very high level of perceived pressure). The Household Crowding Index indicates the number of individuals living in the same house divided by the number of available rooms (Melki et al., 2004).

The second section of the questionnaire contained the Community Assessment of Psychic Experience-42 (CAPE-42) scale, while the third section comprised scales assessing the independent variable (the Columbia Suicide Rating Scale; C-SSRS) and the possible mediator (the Smartphone Addiction Scale-Short Version; SAS-SV).

The CAPE-42 (Konings et al., 2006)

This is a 42-item self-report questionnaire that evaluates positive and negative PEs and depressive symptoms on a two-dimensional scale: (1) frequency of PEs and (2) degree of distress caused by them. For these two dimensions, the total CAPE-42 scores range from 42 to 168. The CAPE-42 has been validated in the Arabic language for use in the current project (Fekih-Romdhane et al., 2023). Only the positive dimension will be used in the present study, and yielded McDonald’s omega value of .78.

The Columbia Suicide Rating Scale (C-SSRS) (Posner et al., 2011)

The C-SSRS is a 5-item measure that assesses the severity of SI during the past week in five different ways: “suicidal thoughts,” “wish to be dead,” “suicidal intent,” “suicidal intent with a specific plan,” and “suicidal thoughts with a method.” Questions were rated as a no/yes type of answer. We used the Arabic version of the scale, which has previously been validated in non-clinical adolescents (Chahine et al., 2020) and young adults (Zakhour et al., 2021). Scores vary between 0 and 5, with 0 indicating the absence of suicidal ideation, whereas higher scores would reflect more severe suicidal ideation. Our sample yielded a McDonald’s omega of .79 for the C-SSRS total scores.

The Smartphone Addiction Scale-Short Version (SAS-SV) (Kwon et al., 2013)

The SAS-SV is a scale comprising of 10 items developed initially in Korea and translated to many languages. This scale is mostly used to confirm the presence of problematic smartphone use in the population and has been previously validated in Arabic (Zeidan et al., 2021). Statements are rated on a 6-point Likert scale, from 1 “strongly disagree” to 6 “strongly agree.” Scores vary from 10 to 66, with higher scores reflecting more smartphone addiction. The McDonald’s omega value for this scale was .92.

Statistical analysis

Data analysis was performed using the SPSS software version 23. We did not have missing values in our database. The positive dimension score had a normal distribution as shown by the skewness and kurtosis values varying between −2 and +2 (George, 2011). Reliability was checked using McDonald’s omega values for the different factors and total scale. The Student t-test was used to compare continuous variables between two groups, and Pearson correlation was used for linear correlation between continuous variables. For categorical variables, the chi-square and Fisher exact tests were used. The PROCESS SPSS Macro version 3.4, model four was used to calculate three pathways (Hayes, 2017). Pathway A determined the effect of suicidal ideation on problematic social media use; Pathway B examined the association between problematic social media use and positive PE, and Pathway C’ estimated the direct effect of suicidal ideation on positive PE, whereas Pathway C estimated the total effect. Significance was deemed present if the Bootstrapped confidence interval of pathway AB did not pass by zero (Hayes, 2017). The mediation analysis was adjusted over all variables that showed a p<0.25 in the bivariate analysis. Significance was set at p < 0.05.

Results

Bivariate analysis

The results of the bivariate analysis are summarized in Table 2. A higher mean positive PEs score was significantly found in single vs married people, in those living alone compared to those living with their family or friends, in those who smoke cigarette, in those who use cannabis or use other illegal drugs, and in those with a history of mental illness. Furthermore, more household crowding index, economic pressure, SI, and PSU were significantly associated with more self-reported positive PEs.

Table 2.

Bivariate analysis of categorical variables associated with the positive CAPE dimension (N=4158)

Categorical variables	Positive PEs scores (Mean ± SD)	P value	Effect size
Gender		.946	.001
Male	33.04 ± 7.40
Female	33.03 ± 7.07
Marital status		.001	.189
Single	33.15 ± 7.23
Married	31.83 ± 6.70
Education		.487	.046
Secondary or less	33.38 ± 8.21
University	33.02 ± 7.14
Residency		.804	.008
Urban	33.00 ± 7.15
Rural	33.06 ± 7.23
Living situation		.005	.050
Alone	34.73 ± 7.60
With family	32.95 ± 7.17
With friends	33.41 ± 5.96
Cigarettes smoking		<.001	.333
No	32.70 ± 6.99
Yes	35.21 ± 8.04
Alcohol drinking		.157	.058
No	32.96 ± 7.08
Yes	33.39 ± 7.71
Cannabis use		.001	.436
No	32.97 ± 7.15
Yes	36.38 ± 8.42
Other illegal drugs use		.011	.457
No	32.99 ± 7.15
Yes	36.89 ± 9.70
Presence of mental illness		<.001	.792
No	31.59 ± 6.27
Yes	37.28 ± 7.99
Continuous variables	r	P value
Suicidal ideation	.38	<.001
Smartphone addiction	.22	<.001
Age	-.02	.302
Household crowding index	.09	<.001
Economic pressure	.12	<.001

Numbers in bold indicate significant p values

Mediation analysis

The results of the mediation analysis are displayed in Table 3. After adjusting for the living situation, cigarette smoking, marital status, household crowding index, economic pressure, cannabis and other illegal drug use, and past history of mental illness, we found that higher levels of suicidal ideation was significantly associated with greater PSU, which was also positively and significantly associated with more positive PEs. Finally, greater suicidal ideation was significantly and directly associated with more positive PEs (Fig. 1).

Table 3.

Mediation analysis: direct and indirect effects of suicidal ideation on positive PEs, taking problematic smartphone use as the mediator

	Direct effect			Indirect effect
	Effect	SE	p	Effect	SE	95% BCa
Suicidal ideation	1.89	0.11	<.001	.10	.02	.06; .15*

*Indicates significant mediation; direct effect= effect of the independent variables on positive PEs in the absence of the mediator; indirect effect=effect of the independent variables on positive PEs in the presence of the mediator; SE, standard error; BCa, bootstrap confidence interval

Fig. 1.

(a) Relation between suicidal ideation and problematic smartphone use (R2= 3.17%); (b) Relation between problematic smartphone use and positive PEs (R2= 23.42%); (c) Total effect of suicidal ideation on positive PEs (R2= 20.59%); (c’) Direct effect of suicidal ideation on positive PEs. Numbers are displayed as regression coefficients (standard error). ***p<0.001

Discussion

Research expanding knowledge on suicide as a factor that precedes the development of psychosis, and might play a role in the emergence and persistence of psychotic symptoms, is highly relevant for clinical and research purposes. Such research may offer novel and promising avenues for the prevention of psychosis in vulnerable individuals. In this context, we considered the possibility that PEs may be the outcome of SI, and we tested the theory-derived hypothesis that PSU mediates this relationship cross-sectionally. Mediation analyses were significant, thus confirming the presence of a significant indirect effect of SI on PEs through PSU.

The present study extends the findings of prior research (Bromet et al., 2017; Connell et al., 2016; Fisher et al., 2013; Honings et al., 2016; Kelleher et al., 2014; Kelleher et al., 2013; Martin et al., 2015), and provides further evidence on the significant positive association between SI and psychotic experiences in non-clinical young adults in a newly explored context, Lebanon. Lebanon is a lower middle income Arab country and has a population mostly composed of adolescents and young people (Chaaban & El Khoury, 2016; du Liban & Unies, 2019). The new Lebanese generations are suffering precarious mental health due to the ongoing crisis (Farran, 2021), including high prevalence of suicidality (Baroud et al., 2019). Currently, early intervention programs have not yet been implemented in Lebanon, and there has been no previous research interest in this area. As such, this is the first study to report the occurrence of PEs and their cross-sectional association with SI in Lebanese young adults from the general population. We, therefore, substantially contribute to the existing body of literature, especially since both SI (An et al., 2020; Chu et al., 2020) and PEs (Vermeiden et al., 2019; Wüsten et al., 2018) are culturally-dependent concepts.

Besides, our study followed the novel framework proposed by Murphy et al. that psychosis may be consequential to SI and behavior (Murphy et al., 2022; Murphy et al., 2018), and attempted to explain parts of the pathway between these variables by examining PSU as a mediator. Thus, another important contribution is the finding that both of the direct and indirect effects of SI on positive PEs are significant, thereby showing a partial mediation of PSU severity in this relationship. Our findings are expected, and in agreement with the existing literature. Regarding the direct effect, the Danish prospective cohort study by Murphy et al. was the first to bring evidence supporting that suicidal behaviors impact psychosis development and persistence (Murphy et al., 2018). Later, Murphy et al. confirmed their hypothesis in a large birth cohort of British twins’ study, where they found that SI and behavior at 12 years significantly increased the risk of PEs at 18 years (Murphy et al., 2022). In terms of the indirect effect, we could find only one previous study that have built on Murphy’s findings, and demonstrated that depressive symptoms partially mediated the path leading from SI to PEs among Chilean high-school adolescents (Nunez et al., 2020). As for PSU, research has shown that healthy young individuals who report general feeling of poor health, and more particularly greater SI, are more likely to overuse their smartphones (Huang et al., 2021; Kim et al., 2019), and be more active on social media than other users (Wang et al., 2019). Besides, digital media use has gained increased attention as the current era’s environmental risk factor for PEs in healthy young adults (Fekih-Romdhane, Sassi, et al., 2021; Lee et al., 2019; McMahon et al., 2021; Mittal et al., 2013). These results could have clinically relevant implications that we discuss below. We draw attention of readers, however, that the cross-sectional design of the study does not allow us to make any causal conclusions. There are plausible reasons to expect that the association between PSU and positive PE may act in one or both directions. Some researchers proposed, for example, that problematic Internet use would serve as a dysfunctional coping strategy to deal with psychotic symptoms, emotional problems, or the often-severe perceived stress characterizing patients with diagnosed psychotic disorder (Chang et al., 2022; Lee et al., 2018). There is also some evidence on bidirectional associations between PSU and SI; with PSU potentially leading to SI in different ways. Individuals who overuse smartphones are at heightened online risks (e.g., cyberbullying) (Durkee et al., 2011; Gansner et al., 2019), increased isolation and decreased social support (Herrero et al., 2019), which may increase, in turn, the vulnerability for suicidality. These observations urge us to interpret the current results with caution, while considering the possible bidirectional interactions between PSU and SI; and call for additional longitudinal studies to confirm our assumptions.

Study limitations

This study has limitations that need to be discussed before drawing any conclusions. The most important limitation lies to its cross-sectional design, which precludes from any causal interpretations. Further longitudinal studies are required to confirm our preliminary findings. In addition, our study was based on self-report measures, which could have led to social desirability and response biases. Also, our sample mostly comprised participants who are single and with a high education level, which might limit the generalizability of our findings. Finally, other variables should be considered subject to further investigation as mediators of the relationship between SI an positive PEs, such as emotional reactivity to stress (Palmier-Claus et al., 2012).

Study implications

Our study demonstrated that PSU played a significant indirect role in the cross-sectional association between SI and positive PEs. This finding suggest that SPU may be regarded as a potential target for the prevention and early intervention in psychosis. Smartphone usage can be viewed either as harmful (Macrynikola et al., 2021) or beneficial (Alao et al., 2006; Sampasa-Kanyinga & Lewis, 2015) for young people with SI. Interestingly, previous longitudinal research documented an improvement in psychotic symptoms and experiences following a discontinuation of digital media usage, while a continued usage was linked to their persistence and worsening (Mittal et al., 2013). As such, clinicians, educators and school administrators ought to give greater concern and attention to PSU among vulnerable young people who present with SI. Patterns of smartphone use should be carefully monitored in suicidal youth, and if necessary, interventions can be proposed for heavy users. Some interventions have proven to be effective in overcoming PSU, such as mindfulness-based cognitive-behavioral therapy (Lan et al., 2018), mind subtraction meditation (Choi et al., 2020), and physical exercise (Liu et al., 2019). Such interventions can act as buffers to prevent psychosis in suicidal individuals with pre-existing constitutional vulnerability. On the other hand, it should be admitted that smartphone use (overuse) has become inevitable in young people lives. As such, clinicians need to rethink their prevention efforts and strategies for PSU by making advantage of the screen time. Findings from limited evidence indicate that smartphone-delivered interventions for individuals wishing to overcome their PSU are acceptable and effective (e.g., (Kent et al., 2021)). However, more research is required in this area before any conclusions can be derived and clear recommendations can be made.

Conclusion

The present study pointed, for the first time, to the mediating role of PSU in the SI/positive PEs connection. Although preliminary, our findings, along with those of previous studies, may help clinicians and researchers gain insight about the mechanisms linking SI to psychosis; and draw attention to potential interventional possibilities. More attention should be given to PSU in suicidal individuals, and interventions should be offered when appropriate. We are aware, however, that additional longitudinal research is still needed to expand our knowledge of the pathways and directionality linking SI, PSU and PEs in healthy young individuals.

Declarations

Conflict of Interest

The authors do not have conflicts of interest to disclose.