The data were collected as part of an ongoing research project examining risk and protective factors for subclinical psychopathology in the Mexican general population. The current sample was recruited between March and July 2022 with individuals who voluntarily agreed to take part in an online survey administered via Qualtrics® software. For the present study, inclusion criteria were being born and currently residing in Mexico and being between 15 and 45 years old. Given the study’s focus on identifying predictors of CHR, exclusion criteria were self-reporting a psychotic spectrum disorder or a psychosis-related hospitalization.

The invitation to participate in the survey was distributed to the general population via personal and institutional (health and educational institutions) social media channels. All participants aged 18 and older provided informed consent. Youth were additionally invited to voluntarily participate through their high schools. School authorities sent information about the purposes of the study to parents or guardians via institutional e-mail and asked them to provide informed consent for their children’s participation. The survey was administered to minors authorized by their parents/guardians to participate and who provided their informed assent. Participants did not receive compensation for completing the survey. The study was approved by the Research Ethics Committee of the Ramón de la Fuente Muñiz National Institute of Psychiatry (CEI/C/019/2021) and conformed to the Helsinki Declaration.

A total of 866 individuals completed the survey. Of these, 38 were excluded for not meeting the inclusion criteria. We also excluded six participants who self-reported a psychotic disorder or a psychosis-related hospitalization. Thus, the analytic dataset included 822 participants.

For sample description, frequencies and percentages were used to summarize categorical variables, and means and standard deviations (S.D) were calculated for the continuous variables. For testing normality, skewness and kurtosis of the variables were calculated. A skewness value over 2 and kurtosis value over 7 are considered to be indicative of moderate or important non-normality (90). Overall, the skewness and kurtosis ranges found in our variables were not out of the norm, with a skewness range from −1.37 to 2.21 and a kurtosis range between −1.98 and 4.76. The only variable that exhibited a higher outside range for skewness (3.92) and kurtosis (13.41) was the item asking whether participants had lived in a place of war/terror or had been exposed to casualties. Given the limited evidence of non-normality, Chi-square tests (χ²) and independent sample Student’s t-tests were used to compare the study variables between those who screened positive for CHR (CHR-positive group) and those who did not (Non-CHR group). The sociodemographic (age, sex, marital/relationship status, educational level, current occupation, and socio-economic status) and psychosocial factors (abuse and neglect during childhood, traumatic events, the risk of presenting problems associated with the use of cannabis, and family functioning) were included as possible predictors associated with a CHR-positive screen in a multivariate logistic regression analysis using the backward conditional modeling approach. To determine the goodness of fit of the models, the Hosmer-Lemeshow test was used (91). Some variables were classified (dummy coded) into auxiliary variables to perform this analysis using their defined cut-off points and were represented by two values, “0” and “1,” where the latter represents the risk/present value. In particular, the variables were coded as follows: marital/relationship status (0 = partnered, 1 = divorced/separated, single), socio-economic level (0 = medium and high, 1 = low), educational level (0 = bachelor studies and higher, 1 = high school studies or less), current occupation (0 = student, economically remunerated activity, 1 = unemployed, non-remunerated activity), cannabis risk level (0 = low risk, 1 = moderate/high risk), family functioning (0 = healthy functioning, i.e., score ≤ 2, 1 = unhealthy functioning, i.e., score > 2), and the previously described cut-offs for clinically significant exposure for the CTQ maltreatment subscales. Two models are presented, the first one containing all the variables in the first step of the modeling approach, and the second including only those variables that remained in the model after the backward stepwise process and which has the most adequate goodness of fit. All tests were deemed to be significant with a p ≤ 0.05. For the analyses, the SPSS version 21 statistical package was used.

The final study sample consisted of 822 individuals. The sample was predominantly female (78.2%, n = 643), with a mean age of 29.1 years (S.D. = 8.3, range 15–45). Most participants were single (55.5%, n = 456; with partner = 40.6%, n = 334; separated/divorced 3.9%, n = 32), with a higher percentage reporting at least a bachelor’s degree (64.4%, n = 529) as well as having an economically remunerated activity (51.9%, n = 427) or being students (35.8%, n = 294). Less than half of the sample self-reported a mental health problem (32.8%, n = 270). Of those with a reported mental health problem, affective (28.1%, n = 76) and anxious symptomatology (27.8%, n = 75) were the most frequently reported. The remaining sociodemographic features are displayed in Table 1.

When dividing the sample into the CHR-positive group and the Non-CHR group, 17.3% (n = 142) met the combined PQ-B and SFQ criteria for a CHR-positive screen. As reported in Table 1, no differences emerged between the groups in terms of sex, marital/relationship status, current occupation, or socio-economic status. Nevertheless, a higher percentage of individuals from the CHR-positive group reported lower educational levels, more self-reported mental health problems, and were younger than those in the Non-CHR group.

Most of the participants reported a low risk of presenting problems associated with cannabis consumption (87.1%, n = 716). However, a higher percentage of CHR-positive individuals reported medium/high risk compared to the Non-CHR group. As displayed in Table 2, in terms of experiences of adversity, no differences arose between groups regarding the experience of living in a place of war/terror (5.5%, n = 45 of the total sample) or experiencing a major natural disaster (20.2%, n = 166 of the total sample). Nevertheless, a higher percentage of those screening positive for CHR experienced a violent or unexpected death of a relative or friend, peer bullying, and intimate partner violence. This group also reported poorer family functioning, more distress associated with the COVID-19 pandemic and lockdown measures, and higher levels of childhood maltreatment across all the CTQ subscales. The differences in childhood maltreatment also emerged when considering the thresholds for the presence of clinically significant levels of exposure. In particular, as compared with the Non-CHR group, the CHR-positive group reported higher rates of emotional abuse (81.7%, n = 116 vs. 48.5%, n = 330; p < 0.001), physical abuse (62.0%, n = 88 vs. 36.6%, n = 249; p < 0.001), sexual abuse (43.7%, n = 62 vs. 25.7%, n = 175; p < 0.001), emotional neglect (49.3%, n = 70 vs. 22.8%, n = 155; p < 0.001), and physical neglect (57.7%, n = 82 vs. 31.2%, n = 212; p < 0.001).

The variables associated with screening positive for CHR for psychosis in the present sample were: having an unhealthy family functioning, a higher risk related to cannabis use, a lower level of completed education, having experienced a major natural disaster, having experienced a violent or unexpected death of a relative or friend, clinically significant levels of emotional abuse, physical neglect, and physical abuse during childhood, and higher distress associated with the COVID-19 pandemic. An older age was a protective factor for screening positive for CHR. The results of the initial and final logistic regression models are displayed in Table 3.

The findings of the comparisons between the CHR-positive and Non-CHR groups showed that individuals screening positive for CHR were younger and reported lower educational levels than those in the Non-CHR group. These differences might be interrelated, as the lower level of highest education attained by the CHR-positive group may correspond to their average younger age. Moreover, the CHR-positive group self-reported more mental health problems than the Non-CHR group, which is consistent with the clinical features of the at-risk stages described in the literature, characterized by the presence of a wide range of non-specific psychopathology and a high prevalence of non-psychotic psychiatric comorbidity, especially with depressive and anxiety disorders (97–100).

The findings regarding differences in the psychosocial domain are in line with previous studies indicating that cannabis misuse, childhood maltreatment, bullying, and traumatic events are more prevalent among CHR individuals, as compared with healthy controls or the general population (32, 37, 47, 67, 101–103). We also found that the CHR-positive group reported poorer family functioning and more distress associated with the COVID-19 pandemic than the Non-CHR group. Although comparisons on these variables are scarcer in the literature, the findings are consistent with the evidence that CHR individuals report worse family functioning or less family satisfaction than Non-CHR individuals (104–106). In addition, they support the results of a recent study suggesting an association between psychotic-like experiences and a greater psychological impact of the COVID-19 pandemic (107). Further research on the effects of the COVID-19 pandemic on this population is needed to confirm this finding and its potential clinical implications.

In line with prior studies (e.g., 10, 40), findings regarding sociodemographic characteristics showed that being older is a protective factor, as it was associated with a lower risk of screening positive for CHR. Furthermore, a lower educational level emerged as a risk factor. Although this supports what has been shown in previous work (10, 44), we interpret this finding with caution, as the most frequent level of education completed by the CHR-positive individuals in our sample was high school, which represents a high or above-average educational level for the Mexican population (108). Therefore, more research in larger and more representative Mexican samples is required to confirm it as a risk factor.

This study found that an unhealthy family functioning was an important psychosocial factor associated with screening positive for CHR for psychosis. This result is notable in light of the centrality of family relationships in the Mexican sociocultural context (109, 110) and seems consistent with the finding that the sociocultural context influences the extent to which components of family functioning are linked to psychotic-like experiences, with a stronger association in developing countries than in highly industrialized countries (60). Future studies directly assessing cultural processes in the Mexican context can enrich our understanding of how culture shapes the interrelations between family factors and psychosis (65, 111, 112). Although family functioning has been less extensively investigated as a predictor of CHR states, existing research on the family environment has indicated that negative family affect and family stress (e.g., criticism, communication problems, less support and cohesion) are associated with poor outcomes across the psychosis continuum, including young people vulnerable to psychosis (55, 56, 58, 113–115), while certain components of a positive family environment (e.g., warmth, positive remarks, family support, cohesion and adaptability, quality of parent–child communication) seem to act as protective factors in the early course of psychosis and have been found to predict improved symptoms or functioning over time in CHR individuals (54, 105, 116–118). Since psychotic symptoms generally emerge during adolescence and early adulthood, a period when individuals are typically greatly impacted by family interactions, these findings suggest the relevance of tailoring early family interventions to bolster family functioning and improve family communication, such as by emphasizing supportive interactions and family psychoeducation (117–120).

Our findings also showed that having a moderate/high risk related to cannabis use was associated with a CHR-positive screen, which is in line with previous research demonstrating an association of cannabis use with psychotic symptoms and psychotic disorders (121–123). Although there is strong evidence indicating a dose–response relationship between cannabis use and psychosis (27, 124, 125), conflicting findings have also emerged (19) and the evidence of the effect of cannabis use on the exacerbation of attenuated psychotic symptoms in CHR individuals is inconsistent (33, 126). Current research on cannabis and psychosis has highlighted several moderating factors, including genetic variability as well as the frequency and age of onset of cannabis use (127, 128). Thus, it will be essential for future research conducted in the Mexican context to examine these factors, especially considering that the COVID-19 pandemic may have impacted the patterns and frequency of cannabis use (129, 130) and that the mean age of onset of cannabis use in Mexico has been decreasing in recent years (131).

Consistent with a robust body of work, we found that early and subsequent exposure to psychosocial stress and adversity was associated with vulnerability to psychosis (12, 27, 30, 37, 132, 133). In particular, findings showed that a CHR-positive screen was associated with having experienced higher levels of emotional abuse, physical abuse, and physical neglect during childhood, a major natural disaster, violent or unexpected death of a relative or friend, and higher distress due to the COVID-19 pandemic.

The association of emotional abuse, physical abuse, and physical neglect with screening positive for CHR parallels previous research demonstrating that a history of childhood maltreatment represents a strong risk factor for psychotic phenomena and the manifestation of the CHR state (19, 37, 39, 42, 47, 134). Furthermore, the findings regarding emotional abuse are in line with a recent review showing that emotional abuse is the type of adversity most strongly associated with subclinical expressions of psychosis in non-clinical samples (135). Therefore, our results concerning childhood maltreatment support those obtained across other populations and further highlight the relevance of distinguishing among maltreatment types when investigating associations with psychosis risk.

Interestingly, some of the other environmental experiences associated with screening positive for CHR in this sample, such as having experienced a major natural disaster and the violent or unexpected death of a relative or friend, have been less prominent in the literature but seem consistent with the particular adverse life experiences of the Mexican population. In recent years, the violence in Mexico has increased substantially, with a historical record of femicides, homicides, and forced disappearances due to organized crime (136). Furthermore, over recent decades, the Mexican population has experienced multiple natural disasters, such as the 2017 earthquakes, as well as floods caused by hurricanes and tropical storms that have resulted in numerous deaths and many victims (damnificados). In addition, although studies are still limited, the daily stressors during the COVID-19 pandemic have impacted the mental health of the Mexican population (137)—and our results suggest that this stress may also be associated with the risk of psychosis. Therefore, further epidemiological and longitudinal research with larger samples is required to better understand the potential medium- and long-term impact of the psychosocial stress and traumatic events identified as risk factors in the present study.

The findings of this study have several clinical implications. The high rate of individuals with a CHR-positive screen in this Mexican sample is alarming and underscores the need to prioritize and promote early detection programs and preventive interventions focused on the early stages of psychosis, which are almost nonexistent in Mexico. Furthermore, if replicated, the psychosocial variables identified as risk factors could help to guide preventive interventions in the population with potential risk for developing psychotic disorders. For instance, both family functioning and cannabis use are amenable to intervention. Therefore, preventive strategies and early interventions could benefit from a focus on reducing dysfunctional family dynamics (54), as well as educational, psychological, and social interventions to reduce cannabis use and its associated risks in CHR adolescents and young adults (31). In addition, the findings support prior calls to make routine and sensitive inquiries about childhood adversity when working with individuals presenting with psychosis-spectrum psychopathology, as well as the importance of implementing effective treatments for childhood trauma in this population (67, 138, 139). Moreover, considering the context of humanitarian emergencies and generalized violence that the Mexican population has experienced, it is necessary to create crisis-focused interventions to reduce the impact of these events on people’s mental health and psychosocial well-being (140).

The current research has certain limitations that should be considered when interpreting the results. First, the cross-sectional nature of the study limits the inferences that can be drawn in terms of causality. Second, all measures were self-reports, which are susceptible to reporting biases. Concerning the designation of a CHR status through self-report measures, we note that we attempted to mitigate this limitation by using strict criteria that included both PQ-B and SFQ cut-offs to establish a CHR-positive screen and that previous work has supported the possibility of detecting CHR individuals through online screening (141). However, it would have been preferable to use widely-validated semi-structured interviews, such as the SIPS (16) or the CAARMS (13). In this regard, it is relevant to note that a recent large-scale study found that both self-report and interview measures of psychotic experiences were associated with similar risk indicators (142). Finally, our use of a non-random sampling technique (participants were individuals with Internet access who were interested in responding to the survey) and the fact that the sample was predominantly female with a relatively high educational level limit the generalizability of the results.

In closing, this study contributes to the body of research examining sociodemographic and psychosocial factors associated with an elevated risk for psychosis across different countries and populations. In particular, we provided, to our knowledge, the first examination of this topic in the Mexican population. Several of the risk factors that emerged as relevant in the present study have been widely reported in the literature, while the significance of other, less extensively studied factors seems consistent with the sociocultural fabric and experiences of the Mexican population. The current work represents a starting place for future studies to build upon, especially by examining these factors longitudinally as well as their interactions with genetic susceptibility. Overall, this study highlights the importance of examining potential contributors to psychosis vulnerability across different sociocultural contexts to delineate risk and protective processes relevant to specific populations and better target preventive intervention efforts.