Abstract
Previous cross-sectional and laboratory research has identified risk factors for persecutory ideation including rumination, negative affect, and safety-seeking behaviors. Questions remain about what in-the-moment factors link general negative affect to PI as well as which maintain PI over time. In the present study, N = 219 individuals completed momentary assessments of PI as well as four factors (attributing threats as certain and important, ruminating, and changing one’s behavior in response) proposed to maintain PI over time. Linear mixed effects models were used to analyze multiple time-varying relationships, including these factors predicting negative affect and vice versa, as well as factors predicting maintenance of PI over time. Linear mixed effects models were used to analyze multiple time-varying relationships, examining each PI-related factor predicting negative affect, negative affect predicting each PI-related factor, as well as each factor predicting maintenance of PI over time. All four factors were associated with increases in subsequent day self-reported severity of PI, suggesting all four increased the likelihood of maintaining or worsening next-day PI. Results of this study confirm that the proposed factors are key in maintaining a cycle by which PI and negative affect are maintained over time. These factors may represent targets for momentary interventions.
Keywords: paranoia, persecutory ideation, mobile health (mHealth), ecological momentary assessment, experience sampling
1. Introduction
Persecutory ideation (PI) describes “unfounded ideas that harm is going to occur and that the persecutor has deliberate intention” (p. 686; Freeman, 2016). These experiences comprise a continuum, ranging from more common and passing worries about threats to highly distressing and disruptive delusions (Bebbington et al., 2013). Most individuals experiencing a first episode of psychosis experience persecutory delusions (Rajapakse et al., 2011) and across presentations, PI is associated with loneliness (Contreras et al., 2022), poor functioning (Freeman et al., 2011) poor well-being (Contreras et al., 2022), anxiety (Freeman et al., 2012), depression (Hartley et al., 2013), and suicidality (Carrillo de Albornoz et al., 2022). While the most severe forms of PI are often associated with schizophrenia-spectrum diagnoses, they also occur in other mental health conditions – for example, bipolar disorder (Smith et al., 2017) or post-traumatic stress disorder (Alsawy et al., 2015) – as well as 2 to 19% of the general population (Freeman et al., 2011). Given its potentially devastating outcomes, conceptual models of PI which lend themselves to empirical testing, are necessary in order to advance our understanding and development of treatment and prevention strategies.
One leading model – the cognitive model of persecutory delusions (Freeman, 2007, 2016; Freeman et al., 2002) – argues that PI is best conceptualized as a “threat belief” that emerges when an individual makes maladaptive interpretations of ambiguous, confusing, or threatening stimuli. Threat beliefs are maintained (i.e. continue to be salient moment-to-moment) when negative affective states reduce cognitive resources required to revise initial interpretations. Negative affect comprises internal states associated with both non-reward and anticipatory concerns (American Psychological Association, 2018) and often emerges in the context of symptoms of depression, anxiety or distress. According to the cognitive model, threat beliefs are furthered when individuals avoid situations that would challenge such beliefs (e.g., by staying inside to avoid a believed conspiracy, the individual receives no information that would contradict the initial belief), a phenomenon referred to as safety seeking. This model has been supported by early studies demonstrating cross-sectional relationships of a number of factors to PI, including worry or rumination (Hartley et al., 2014; O’Driscoll et al., 2014), negative self-beliefs (Kesting & Lincoln, 2013), and safety-seeking behaviors (Freeman et al., 2007).
Early studies examining predictors of PI primarily relied upon cross-sectional or experimental laboratory research designs and retrospective measures. In addition to being susceptible to memory errors, biases in reporting, or demand characteristics, retrospective measures do not provide granular information about the real-time antecedents and consequences of PI. More recently, technology-assisted approaches – e.g. ecological momentary assessments (EMA) or experience sampling methodology (ESM) – have provided new opportunities to address these limitations. Studies using PDAs (Ben-Zeev et al., 2011), pagers (with accompanying journals; Thewissen et al., 2008, 2011) and later smartphones (Kramer et al., 2014) have provided support for components of the cognitive model of PI. Ludtke and colleagues (2023) summarized this literature in a recent systematic review finding strong evidence supporting two predictors: negative affect and sleep problems. Their review also identified a number of limitations. First, though PI ranges in severity cross diagnostically, most longitudinal studies have examined PI in either individuals diagnosed with schizophrenia-spectrum disorders or amongst non-clinical controls. This underrepresents individuals with mild or moderate PI and could impede identifying attributes that lead to significant distress and disruption across the continuum (McGrath et al., 2015; Van Os & Reininghaus, 2016). Second, a minority of studies have modeled within-participant effects (i.e. controlling for participant mean values), which are the standard of evidence for identifying each factor as a precursor to PI. In Ludtke’s review (2023), authors could not judge whether time-lagged and within-person associations existed for these factors given insufficient data. Most studies have examined affective or behavioral characteristics that increase the likelihood of PI; fewer studies have examined the impact of responses to PI once it emerges.
Cognitive approaches to persecutory ideation are designed to challenge the certainty one feels about a particular threat (Freeman, 2013); for example, challenging the veracity of intrusive worries that one is being spied on by a neighbor. Approaches built on the principles of acceptance and commitment therapy (ACT) work toward intentional action consistent with one’s values in place of preoccupation with one’s threat beliefs; for example, persisting in meaningful or required daily activity even while being distressed. While several studies (Bloy et al., 2011; Garety et al., 2021; Johns et al., 2016) have provided evidence to support these approaches, few studies have examined changes in theory-derived intervention targets on a moment-to-moment time scale. Understanding momentary predictors can help identify targets for momentary interventions, for example, just-in-time adaptive interventions provided through digital platforms.
Our team has developed and deployed a smartphone data collection system designed for longitudinal studies and deployed that system in samples of individuals with schizophrenia (Ben-Zeev et al., 2017; Buck et al., 2019; Wang et al., 2016) and cross-diagnostic auditory verbal hallucinations or voices (Ben-Zeev et al., 2020). More recently, we recruited a sample of over 200 individuals with cross-diagnostic PI to carry and respond to ecological momentary assessment (EMA) questionnaires over a 30-day period (Buck et al., 2023). The present study involves an examination of four factors proposed to maintain (i.e. sustain or worsen once emerged) persecutory ideation once it emerges: appraisals of certainty about the threat, appraisals of its importance, rumination, and changing one’s behavior in response. Specifically, we aimed to determine the extent to which each of these factors are associated with negative affect both between-participants (i.e. at the mean level) and in time-lagged within-participants analyses, and whether each predicts whether PI will be maintained over time within-participants (i.e. on the day-to-day level). In order to assess the impact of these factors on a more sustained and clinically meaningful time scale, we assessed changes that occurred on a day-to-day level. Based on the cognitive model (Freeman et al., 2002), we propose that that all four proposed factors – attributing threats as certain and important, ruminating about them, and changing one’s behavior in response – will both be associated with increases in previous negative affect and serve as precursors to worsening negative affect and persecutory ideation. This would support their role as key links in cycles that maintain paranoia over time and support momentary interventions to address them.
2. Methods
2.1. Participants.
Inclusion criteria included: (1) experiencing persecutory ideation as defined by a score of 11 or greater on the Revised Green Paranoid Thoughts Scale – Persecution Subscale, (2) being 18 years or older, (3) being a native English speaker, (4) owning and willing to use an Android smartphone (with an active data plan) for study data collection. Candidates were excluded if they (1) lived outside of the US, (2) expected to be unavailable for any part of the 30-day data collection period, or (3) had previously participated. Participants were recruited entirely online through Google Ads and email listservs of previous study completers who had interacted with our group. This study includes 219 participants who provided EMA responses items included in analysis of momentary changes.
2.2. Smartphone data collection.
Data were collected using a multimodal smartphone data collection. Participants were asked to install this system on their device, carry it for 30 days and respond to brief surveys in real-time (i.e. EMAs) for 30 days. They were prompted to complete questionnaires four times daily, semi-randomly between 9AM and 9PM. Though entries could also be manually initiated by participants, in order to avoid a biased sample of responses only prompted ones are analyzed in this study. EMA items asked participants to report on their current experiences assessed on a 4-point Likert scale (0 = not at all, 3 = very much). A complete description of this system can be found in our team’s related manuscript (Buck et al., 2023); six items were examined in detail in the current manuscript. First, participants were asked about PI occurrence (“Are you currently feeling that someone or something wants to harm you?”). If endorsed above a 0 (“not at all”) participants were asked additional items about their current experiences. Three of these items assessed cognitive responses including (1) conviction (“How certain are you that this is true?”), (2) threat importance (“How much does the threat matter to you?”), and (3) rumination (“How much are you thinking about the threat?”), and the fourth assessed behavior change (“How much are you changing your behavior right now because of the threat?”). Additional items assessed sadness (“How sad do you feel right now?”) and anxiety (“How anxious do you feel right now?”); following our group’s previous work (Buck et al., 2022), we combined sadness and anxiety to assess negative affect in one variable; each of these variables were highly correlated with one another (r = .64). Sadness and anxiety were used to approximate negative affective states broadly given evidence in this literature for the role of negative affect broadly (Lüdtke et al., 2023) as well as definitions of negative affect that comprise both non-reward and anticipatory concerns (American Psychological Association, 2018).
2.3. Procedures.
Study procedures were based on our team’s previous work examining auditory verbal hallucinations (Ben-Zeev et al., 2020). Full procedures have also been reported elsewhere in our group’s manuscript examining associations between collected variables and functioning (Buck et al., 2023). Participants were recruited remotely using Google Ads that were optimized to appear in prospective participants’ search results based on specific search keywords. Ads would appear following search queries consistent with medical (e.g., schizophrenia, bipolar) or non-medical descriptions of illness (e.g., being followed, conspiracy), or if the search terms matched the automatically generated “broad match” keywords generated through the Google Ads system. Clicking on an ad link brought candidates to the study website where information was provided about the study and smartphone data collection system. The website also included a link to the screening and consent questionnaire, where interested participants could verify their phone number and email address, review the study consent form, and complete all screening questionnaires. Eligible participants could then immediately provide informed consent, complete baseline questionnaires, and then were provided access and instructions to install the study mobile application.
During the 30-day data collection period, participants were instructed to carry their smartphone devices with them and respond to prompts. Participants were encouraged to reach out to the research team during the data collection period for technical assistance. The research team would call or text a participant if their device did not collect data for three days. At the end of the 30-day period, the application would cease data collection and participants were asked to uninstall the app. Participants were compensated $75 for participation and $50 to cover any additional data charges incurred for a total of $125.
2.4. Data analytic plan.
We examined the relationships of momentary appraisals to concurrent and prospective affect and behavior using linear mixed effects models, using IBM SPSS Statistics Version 28. All available data were used for each model. Models that examined the relationships of PI-related appraisals to measures of affect and behavior assessed through EMA followed a parallel structure, in an approach we first validated in work examining the relationship of appraisals to affect and behavior in a sample of individuals experiencing auditory verbal hallucinations or voices (Buck et al., 2022). In all models, outcomes included either persecutory ideation, negative affect, or each factor (certainty, importance, rumination, and behavior change). We used multiple terms to represent different temporal elements of these relationships. The first term was the participant’s mean of the predictor variable to examine between-participant variance. This model term represents the association of the participant average (across the full data collection period) to the participant’s average of the modeled outcome. A second term is the participant mean-centered time-lagged daily rating. This variable was derived through subtracting the average score of the predictor on a particular day from the participant mean during the entire study period. This variable thus represents the previous day value of the predictor entirely separate from the influence of the individual’s average score on the predictor. For models examining maintenance (sections 3.2 and 3.3; rather than just within-subject associations), we also generated time-lagged, mean-centered values of each outcome to allow for models that examined prediction of change in the outcome from one day to the next rather than elevation relative to the participant mean. This approach – which has been employed in a minority of studies examining temporal changes in PI (Lüdtke et al., 2023) – allows not only for examination of within-participant change, but also whether an increase in one variable predicts change in another from one time point to the next.
3. Results
Participant demographics can be found in Table 1. Additional clinical details as well as information related to system engagement are reported elsewhere (Buck et al., 2023).
Table 1.
Sample descriptive statistics.
| M or N | (SD or %) | |
|---|---|---|
| Age | 38.21 | (11.64) |
| Race | ||
| White / Caucasian | 154 | (70.3%) |
| Black / African American | 39 | (17.8%) |
| Pacific Islander | 1 | (0.5%) |
| American Indian / Alaskan Native | 2 | (0.9%) |
| Asian | 5 | (2.3%) |
| More than one race | 18 | (8.2%) |
| Gender | ||
| Female | 155 | (70.8%) |
| Male | 54 | (24.7%) |
| Transgender man | 3 | (1.4%) |
| Transgender woman | 2 | (0.9%) |
| Other / Non-binary | 5 | (2.3%) |
| Ethnicity | ||
| Non-Hispanic / Non-Latino | 204 | (93.2%) |
| Hispanic / Latino | 15 | (6.8%) |
| Diagnoses (self-report) | ||
| Alzheimer’s or Parkinson’s | 2 | (0.9%) |
| Bipolar disorder | 90 | (41.5%) |
| Depression | 146 | (67.3%) |
| Borderline Personality Disorder | 36 | (16.6%) |
| Paranoid Personality Disorder | 16 | (7.4%) |
| Schizoaffective Disorder | 33 | (15.2%) |
| Schizophrenia | 31 | (14.3%) |
| Post Traumatic Stress Disorder | 110 | (50.7%) |
| Substance Use | 55 | (25.3%) |
| Schizotypal Personality Disorder | 3 | (1.4%) |
| Anxiety Disorder | 141 | (65.0%) |
| None | 26 | (12.0%) |
3.1. Associations between negative affect and PI-related factors.
A first set of models examined the relationship of negative affect to the proposed factors associated with PI (see Table 2). First, between-participant mean values of negative affect were associated with elevated averages in all four factors, including conviction (B = 0.57, SE = 0.07, p < .001), importance (B = 0.46, SE = 0.06, p < .001), rumination (B = 0.54, SE = 0.06, p < .001) and behavior change (B = 0.57, SE = 0.07, p < .001). This suggests that individuals who are more likely to attribute certainty and importance to their threats, as well as those who ruminate and change their behavior in response are also more likely to experience increased negative affect. Second, models also revealed that within-participants, elevations in negative affect were associated with increases in next-day values in each of the four factors: conviction (B = 0.13, SE = 0.02, p < .001), importance (B = 0.13, SE = 0.03, p < .001), rumination (B = 0.11, SE = 0.02, p < .001) and behavior change (B = 0.13, SE = 0.02, p < .001). This suggests that within individual change in negative affect was associated with elevations the following day in all factors relative to the participant’s average. We then examined whether the reverse was true, namely, whether PI-related factors were associated with prospective negative affect, controlling for the participant’s mean value of each factor. This was true for all four, as mean-centered conviction (B = 0.08, SE = 0.02, p < .001), importance (B = 0.11, SE = 0.02, p < .001), rumination (B = 0.11, SE = 0.02, p < .001) and behavior change (B = 0.08, SE = 0.02, p < .001) were all associated with next-day negative affect.
Table 2.
Associations of negative affect to next-day PI-related factors.
| How certain are you that this is true? | How much does the threat matter to you? | |||||||
|---|---|---|---|---|---|---|---|---|
| Dependent variable: Next day conviction | Est | SE | t / Wald Z | Dependent variable: Next day importance | Est | SE | t | |
| Fixed effects | Fixed effects | |||||||
| Intercept | 1.04 | 0.11 | 9.86*** | Intercept | 1.41 | 0.09 | 14.95*** | |
| Negative Affect – Mean | 0.57 | 0.07 | 8.35*** | Negative Affect – Mean | 0.46 | 0.06 | 7.67*** | |
| Negative Affect – Previous Day | 0.13 | 0.02 | 6.34*** | Negative Affect – Previous Day | 0.13 | 0.03 | 6.26*** | |
| Random effects | Random effects | |||||||
| Residual | 0.36 | 0.01 | 35.95*** | Residual | 0.36 | 0.01 | 35.91*** | |
| Intercept (subject) | 0.37 | 0.04 | 9.01*** | Intercept (subject) | 0.28 | 0.03 | 8.47*** | |
| Observations | 2808 | Observations | 2815 | |||||
| AIC | 5652.62 | AIC | 5623.45 | |||||
| Rumination | Behavior Change | |||||||
| How much are you thinking about the threat? | How much are you changing your behavior in response? | |||||||
| Dependent variable: Next day rumination | Est | SE | t / Wald Z | Dependent variable: Next day behavior change | Est | SE | t | |
| Fixed effects | Fixed effects | |||||||
| Intercept | 1.16 | 0.09 | 13.33*** | Intercept | 0.64 | 0.10 | 6.18*** | |
| Negative Affect – Mean | 0.54 | 0.06 | 9.81*** | Negative Affect – Mean | 0.57 | 0.07 | 8.64*** | |
| Negative Affect – Previous Day | 0.11 | 0.02 | 5.22*** | Negative Affect – Previous Day | 0.13 | 0.02 | 5.25*** | |
| Random effects | Random effects | |||||||
| Residual | 0.37 | 0.01 | 36.03*** | Residual | 0.48 | 0.01 | 36.13*** | |
| Intercept (subject) | 0.23 | 0.03 | 8.39*** | Intercept (subject) | 0.33 | 0.04 | 8.85*** | |
| Observations | 2822 | 2821 | ||||||
| AIC | 5673.62 | 6392.50 | ||||||
3.2. Maintenance or increase of prospective negative affect.
A second set of models examined the relationships of the same factors to prospective negative affect controlling for the previous day’s negative affect (see Table 3). These models revealed that above and beyond the previous day’s negative affect and the participant’s mean value of the PI-related factor, both importance (B = 0.04, SE = 0.02, p = .03) and rumination (B = 0.04, SE = 0.02, p = .02) were associated with next-day increases in negative affect, but the same was not true for conviction or behavior change (p values > .14). This suggests that on days when individuals experienced PI, their tendency to appraise those threats as important and ruminate about them were associated with increases in negative affect the subsequent day. In other words, these factors appeared to maintain or worsen already present negative affect.
Table 3.
PI-related factors predicting change in next-day negative affect.
| How certain are you that this is true? | How much does the threat matter to you? | ||||||
|---|---|---|---|---|---|---|---|
| Dependent variable: Next day negative affect | Est | SE | t / Wald Z | Dependent variable: Next day negative affect | Est | SE | t |
| Fixed effects | Fixed effects | ||||||
| Intercept | 0.53 | 0.11 | 4.67*** | Intercept | 0.39 | 0.15 | 2.66** |
| Negative Affect – Previous day | 0.24 | 0.02 | 11.72*** | Negative Affect – Previous day | 0.23 | 0.02 | 11.24*** |
| Conviction – Mean | 0.47 | 0.06 | 8.25*** | Importance – Mean | 0.49 | 0.07 | 7.30*** |
| Conviction – Previous day | 0.03 | 0.02 | 1.16 | Importance – Previous day | 0.04 | 0.02 | 2.20* |
| Random effects | Random effects | ||||||
| Residual | 0.30 | 0.01 | 36.34*** | Residual | 0.30 | 0.01 | 36.37*** |
| Intercept (subject) | 0.33 | 0.04 | 9.15*** | Intercept (subject) | 0.35 | 0.04 | 9.26*** |
| Observations | 2864 | Observations | 2867 | ||||
| AIC | 5272.85 | AIC | 5268.97 | ||||
| Rumination | Behavior Change | ||||||
| How much are you thinking about the threat? | How much are you changing your behavior in response? | ||||||
| Dependent variable: Next day negative affect | Est | SE | t / Wald Z | Dependent variable: Next day negative affect | Est | SE | t |
| Fixed effects | Fixed effects | ||||||
| Intercept | 0.23 | 0.13 | 1.78^ | Intercept | 0.67 | 0.10 | 6.99*** |
| Negative Affect – Previous day | 0.23 | 0.02 | 10.96*** | Negative Affect – Previous day | 0.23 | 0.02 | 11.44*** |
| Rumination – Mean | 0.60 | 0.06 | 9.29*** | Behavior Change – Mean | 0.50 | 0.06 | 8.50*** |
| Rumination – Previous day | 0.04 | 0.02 | 2.36* | Behavior Change – Previous day | 0.02 | 0.02 | 1.47 |
| Random effects | Random effects | ||||||
| Residual | 0.30 | 0.01 | 36.41*** | Residual | 0.30 | 0.01 | 36.41*** |
| Intercept (subject) | 0.30 | 0.03 | 9.11*** | Intercept (subject) | 0.32 | 0.03 | 9.16*** |
| Observations | 2875 | 2875 | |||||
| AIC | 5277.81 | 5291.30 | |||||
3.3. Maintenance or increase of prospective persecutory ideation.
Our final models examined whether each predictor was associated with next-day increases in PI above and beyond concurrent levels, i.e. whether each indeed maintained PI over time (see Table 4). While controlling for the previous day’s persecutory ideation score and the participant’s mean score on each factor, each was significantly associated with next day increases in persecutory ideation. This applied both to appraisals that persecutory threats were certainly true (B = 0.09, SE = 0.02, p < .001) and that the threat was important (B = 0.08, SE = 0.02, p < .001), as well as reporting ruminating about the threat (B = 0.06, SE = 0.02, p = .001) and changing one’s behavior because of the threat (B = 0.05, SE = 0.02, p < .001). In other words, on days when individuals were appraising their threats as more certain, more important, ruminating about them, or changing behavior in response to them, the next day they experienced greater PI. This supports the hypothesis that these four attributes can be regarded as factors that maintain PI over time.
Table 4.
PI-related factors predicting change in next-day persecutory ideation.
| How certain are you that this is true? | How much does the threat matter to you? | ||||||
|---|---|---|---|---|---|---|---|
| Dependent variable: Next day persecutory ideation | Est | SE | t / Wald Z | Dependent variable: Next day persecutory ideation | Est | SE | t |
| Fixed effects | Fixed effects | ||||||
| Intercept | −0.38 | 0.14 | −2.73** | Intercept | −0.50 | −0.19 | −2.65*** |
| Persecutory ideation – Previous day | 0.19 | 0.02 | 8.59*** | Persecutory ideation – Previous day | 0.20 | 0.02 | 9.45*** |
| Conviction – Mean | 0.75 | 0.07 | 10.68*** | Importance – Mean | 0.72 | 0.09 | 8.34*** |
| Conviction – Previous day | 0.09 | 0.02 | 4.61*** | Importance – Previous day | 0.08 | 0.02 | 3.84*** |
| Random effects | Random effects | ||||||
| Residual | 0.34 | 0.01 | 36.57*** | Residual | 0.34 | 0.01 | 36.61*** |
| Intercept (subject) | 0.53 | 0.06 | 9.45*** | Intercept (subject) | 0.62 | 0.06 | 9.66*** |
| Observations | 2899 | Observations | 2902 | ||||
| AIC | 5711.30 | AIC | 5734.46 | ||||
| Rumination | Behavior Change | ||||||
| How much are you thinking about the threat? | How much are you changing your behavior in response? | ||||||
| Dependent variable: Next day persecutory ideation | Est | SE | t / Wald Z | Dependent variable: Next day persecutory ideation | Est | SE | t |
| Fixed effects | Fixed effects | ||||||
| Intercept | −0.56 | 0.18 | −3.17** | Intercept | −0.08 | 0.13 | −0.64 |
| Persecutory ideation – Previous day | 0.20 | 0.02 | 9.34*** | Persecutory ideation – Previous day | 0.20 | 0.02 | 9.54*** |
| Rumination – Mean | 0.81 | 0.09 | 9.27*** | Behavior Change – Mean | 0.74 | 0.08 | 9.61*** |
| Rumination – Previous day | 0.06 | 0.02 | 3.27** | Behavior Change – Previous day | 0.05 | 0.02 | 3.30*** |
| Random effects | Random effects | ||||||
| Residual | 0.34 | 0.01 | 36.66*** | Residual | 0.34 | 0.01 | 36.66*** |
| Intercept (subject) | 0.59 | 0.06 | 9.63*** | Intercept (subject) | 0.58 | 0.06 | 9.67*** |
| Observations | 2909 | 2907 | |||||
| AIC | 5747.04 | 5724.06 | |||||
4. Discussion
The present study provides granularity to the examination of how PI is maintained over time. We tested four factors proposed to maintain persecutory ideation: attributing threats as certain and important, ruminating about them, and changing one’s behavior in response. First, individuals that experience elevated negative affect are more likely to respond to thoughts about potential threats in maladaptive ways, and when distressed are even more likely to do so. Second, once PI emerges, elevations in each of these four factors make it more likely that negative affect or PI will be worse the subsequent day. Models supported a “vicious cycle” model of PI wherein distress worsens maladaptive responses to threat beliefs, and maladaptive responses to threat beliefs worsens distress. These results identify four potentially modifiable factors that are closely associated with real-time maintenance or increase of persecutory ideation and thus support interventions designed to disrupt each of them.
Broadly, data provide support for the cognitive model of PI. The cognitive model asserts PI stems from distorted appraisals of uncertain situations resulting in increased perceptions of threats (Freeman, 2007). Results align with previous work highlighting factors that increase these appraisals and that these appraisals have specific consequences on future affect, thought, and behavior. This study supports and extends the work of Ben-Zeev and colleagues (2011) that supported a link between anxiety strengthen future belief conviction. We build upon this work by showing how this subsequent effect on belief conviction goes on to further exacerbate experiences of negative affect and increase the likelihood of PI; this study also extends this pattern to other PI-related factors, including behavior change, rumination, and a largely unstudied construct, threat importance. The use of a real-time, real-place assessment system over an extended period allows this manuscript to find evidence for maintenance cycles that unfold in real-time and increase negative impacts of PI.
This study has important potential implications for providing therapeutic intervention to people experiencing PI. It supports the use of cognitive interventions (e.g., CBT for psychosis, Acceptance and Commitment Therapy, Cognitive Restructuring) which target factors that worsen psychotic symptoms in the moment. Specifically, it provides clinicians insight about the importance of providing skills to intervene on threat certainty and importance, ruminations about perceptions, and behavior impulses that result from these perceptions. By specifically providing cognitive restructuring intervention on beliefs that threats are likely to happen or important, individuals experiencing PI may begin to view events as less imminent and less harmful. Use of distraction techniques, behavioral activation, and engaging with meaningful goals may help individuals reduce rumination, avoidance, and reliance on “safety seeking” behaviors as well.
Beyond confirming the mechanisms by which current popular cognitive interventions are likely to be helpful for people with PI, this study also highlights the potential impact of ecological momentary interventions (EMIs) for PI (Myin-Germeys et al., 2016). Notably, our results revealed both between- and within-individual associations between PI-related factors and distress. This suggests ongoing real-time intervention on these factors could reduce distress and reduce PI over time. EMIs leverage mobile devices to provide real-time interventions to people based on responses to data indicating an in-the-moment need. Data for this study was collected in a real-time, real-place manner similar to how an EMI would gather data prior to providing support. Through the use of EMI, it may be possible to disrupt cycles more rapidly than what happens in traditional psychotherapy and provide support more efficiently. One example of an effective EMI for PI is called SlowMo, which uses a website to teach individuals how they can intervene on the thoughts and feelings associated with PI. Following, the intervention uses a mobile app to check-in with question prompts and provide reminders to use skills based on prompt responses (Garety et al., 2021). Evidence suggests that users of this intervention saw improvement in 12-week ratings of paranoia outcomes, including the specific feature of belief flexibility.
The present study is not without several notable limitations. First, the study was conducted online, required an Android smartphone, and was conducted with a sample with an overrepresentation of participants that identify as female. Because the study was only available to people who had access to a smartphone capable of running our data collection program (i.e. an Android device), results may not be representative of the full range of people who experience PI. The gender representation of our study is consistent with other studies conducted online with samples of people with symptoms of mental illness (Ben-Zeev et al., 2021). Future studies can examine whether there are systematic biases to the ways in which online recruitment engages participants that identify as women, and the extent to which these biases resemble those in mental health services broadly. Another limitation of the study is that we did not collect data on abnormal sensory experiences or reasoning biases. These are two often cited factors (Freeman, 2016) associated with PI and we are unable to comment on how these factors intersect with the constructs studied in the present work.
Research increasingly supports that there are multiple, detectable risk and maintenance factors that perpetuate the cycle of PI. By identifying these factors, it becomes possible to design treatment interventions to specifically target them with a goal of interrupting the development of PI. The present study confirms previous findings about particular PI-related factors and extends it by using data collected using real-time, real-place methodology. We found that appraisal of threat importance plays a specific role in furthering future negative affect and PI. Continued research is needed to develop clinically useful interventions based on these findings and study whether intervening on these specific factors may break the vicious cycle of PI.
Highlights.
Momentary responses to persecutory ideation predict maintenance of PI
Conviction, importance, rumination, and behavior change are linked with negative affect
These factors both increase distress and are worsened by it prospectively
These maintenance factors may be a target for effective ecological momentary interventions
Role of the funding source
Research reported in this manuscript was supported by a grant from the National Institute of Mental Health (NIMH) under award numbers R01MH112641 as well as a NARSAD Young Investigator Grant from the Brain and Behavior Research Foundation (Grant #27917). Dr. Buck is also supported by a Mentored Patient-Oriented Research Career Development Award from NIMH (K23MH122504). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Brain and Behavior Research Foundation. The organizations supporting this research had no role in the design, collection, analysis, interpretation or writing of the manuscript.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of competing interest
Dr. Ben-Zeev has financial interests in FOCUS technology, CORE technology, and Merlin LLC. He has an intervention content licensing agreement with Pear Therapeutics and has provided consultation services to Trusst Health, K Health, Boehringer Ingelheim, eQuility, Deep Valley Labs, and Otsuka Pharmaceuticals. The other authors have no conflicts to disclose.
References
- Alsawy S, Wood L, Taylor PJ, & Morrison AP (2015). Psychotic experiences and PTSD: Exploring associations in a population survey. Psychological Medicine, 45(13), 2849–2859. 10.1017/S003329171500080X [DOI] [PubMed] [Google Scholar]
- American Psychological Association. (2018). Negative affect. APA Dictionary of Psychology. https://dictionary.apa.org/negative-affect [Google Scholar]
- Bebbington PE, McBride O, Steel C, Kuipers E, Radovanoviĉ M, Brugha T, Jenkins R, Meltzer HI, & Freeman D (2013). The structure of paranoia in the general population. British Journal of Psychiatry, 202(6), 419–427. 10.1192/bjp.bp.112.119032 [DOI] [PubMed] [Google Scholar]
- Ben-Zeev D, Brian R, Wang R, Wang W, Campbell AT, Aung MSH, Merrill M, Tseng VWS, Choudhury T, Hauser M, Kane JM, & Scherer EA (2017). CrossCheck: Integrating self-report, behavioral sensing, and smartphone use to identify digital indicators of psychotic relapse. Psychiatric Rehabilitation Journal, 40(3), 266–275. 10.1037/prj0000243 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ben-Zeev D, Buck B, Chander A, Brian R, Wang W, Atkins D, Brenner CJ, Cohen T, Campbell A, & Munson J (2020). Mobile RDoC: Using Smartphones to Understand the Relationship Between Auditory Verbal Hallucinations and Need for Care. Schizophrenia Bulletin Open, 1(1), sgaa060. 10.1093/schizbullopen/sgaa060 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ben-Zeev D, Chander A, Tauscher J, Buck B, Nepal S, Campbell A, & Doron G (2021). A Smartphone Intervention for People With Serious Mental Illness: Fully Remote Randomized Controlled Trial of CORE. Journal of Medical Internet Research, 23(11), e29201. 10.2196/29201 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ben-Zeev D, Ellington K, Swendsen J, & Granholm E (2011). Examining a cognitive model of persecutory ideation in the daily life of people with schizophrenia: A computerized experience sampling study. Schizophrenia Bulletin, 37(6), 1248–1256. 10.1093/schbul/sbq041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bloy S, Oliver JE, & Morris E (2011). Using Acceptance and Commitment Therapy With People With Psychosis. Clinical Case Studies, 10(5), 347–359. 10.1177/1534650111420863 [DOI] [Google Scholar]
- Buck B, Munson J, Chander A, Wang W, Brenner CJ, Campbell AT, & Ben-Zeev D (2022). The relationship between appraisals of auditory verbal hallucinations and real-time affect and social functioning. Schizophrenia Research, 250, 112–119. 10.1016/j.schres.2022.10.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Buck B, Scherer E, Brian R, Wang R, Wang W, Campbell A, Choudhury T, Hauser M, Kane JM, & Ben-Zeev D (2019). Relationships between smartphone social behavior and relapse in schizophrenia: A preliminary report. Schizophrenia Research, 208, 167–172. 10.1016/j.schres.2019.03.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Buck B, Wingerson M, Tauscher JS, Enkema M, Wang W, Campbell AT, & Ben-Zeev D (2023). Using Smartphones to Identify Momentary Characteristics of Persecutory Ideation Associated With Functional Disability. Schizophrenia Bulletin Open, 4(1). 10.1093/schizbullopen/sgad021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carrillo de Albornoz CM, Gutiérrez B, Ibanez-Casas I, & Cervilla JA (2022). Paranoia and Suicidality: A Cross-Sectional Study in the General Population. Archives of Suicide Research, 26(3), 1587–1599. 10.1080/13811118.2021.1950589 [DOI] [PubMed] [Google Scholar]
- Contreras A, Valiente C, Vázquez C, Trucharte A, Peinado V, Varese F, & Bentall RP (2022). The network structure of paranoia dimensions and its mental health correlates in the general population: The core role of loneliness. Schizophrenia Research, 246, 65–73. 10.1016/j.schres.2022.06.005 [DOI] [PubMed] [Google Scholar]
- Freeman D (2007). Suspicious minds: The psychology of persecutory delusions. Clinical Psychology Review, 27(4), 425–457. 10.1016/j.cpr.2006.10.004 [DOI] [PubMed] [Google Scholar]
- Freeman D (2013). Cognitive-Behavioral Therapy for Psychotic Disorders. Psychiatric Times, 30(12). [Google Scholar]
- Freeman D (2016). Persecutory delusions: a cognitive perspective on understanding and treatment. The Lancet Psychiatry, 3(7), 685–692. 10.1016/S2215-0366(16)00066-3 [DOI] [PubMed] [Google Scholar]
- Freeman D, Garety PA, Kuipers E, Fowler D, & Bebbington PE (2002). A cognitive model of persecutory delusions. British Journal of Clinical Psychology, 41(4), 331–347. [DOI] [PubMed] [Google Scholar]
- Freeman D, Garety PA, Kuipers E, Fowler D, Bebbington PE, & Dunn G (2007). Acting on persecutory delusions: The importance of safety seeking. Behaviour Research and Therapy, 45(1), 89–99. 10.1016/j.brat.2006.01.014 [DOI] [PubMed] [Google Scholar]
- Freeman D, McManus S, Brugha T, Meltzer H, Jenkins R, & Bebbington P (2011). Concomitants of paranoia in the general population. Psychological Medicine, 41(05), 923–936. 10.1017/S0033291710001546 [DOI] [PubMed] [Google Scholar]
- Freeman D, Stahl D, McManus S, Meltzer H, Brugha T, Wiles N, & Bebbington P (2012). Insomnia, worry, anxiety and depression as predictors of the occurrence and persistence of paranoid thinking. Social Psychiatry and Psychiatric Epidemiology, 47(8), 1195–1203. 10.1007/s00127-011-0433-1 [DOI] [PubMed] [Google Scholar]
- Garety P, Ward T, Emsley R, Greenwood K, Freeman D, Fowler D, Kuipers E, Bebbington P, Dunn G, & Hardy A (2021). Digitally supported CBT to reduce paranoia and improve reasoning for people with schizophrenia-spectrum psychosis: the SlowMo RCT. Efficacy and Mechanism Evaluation, 8(11). 10.3310/eme08110 [DOI] [PubMed] [Google Scholar]
- Garety P, Ward T, Emsley R, Greenwood K, Freeman D, Fowler D, Kuipers E, Bebbington P, Rus-Calafell M, McGourty A, Sacadura C, Collett N, James K, & Hardy A (2021). Effects of SlowMo, a Blended Digital Therapy Targeting Reasoning, on Paranoia Among People With Psychosis. JAMA Psychiatry, 78(7), 714. 10.1001/jamapsychiatry.2021.0326 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hartley S, Barrowclough C, & Haddock G (2013). Anxiety and depression in psychosis: a systematic review of associations with positive psychotic symptoms. Acta Psychiatrica Scandinavica, 128(5), 327–346. 10.1111/acps.12080 [DOI] [PubMed] [Google Scholar]
- Hartley S, Haddock G, Vasconcelos e Sa D, Emsley R, & Barrowclough C (2014). An experience sampling study of worry and rumination in psychosis. Psychological Medicine, 44(8), 1605–1614. 10.1017/S0033291713002080 [DOI] [PubMed] [Google Scholar]
- Johns LC, Oliver JE, Khondoker M, Byrne M, Jolley S, Wykes T, Joseph C, Butler L, Craig T, & Morris EMJ (2016). The feasibility and acceptability of a brief Acceptance and Commitment Therapy (ACT) group intervention for people with psychosis: The ‘ACT for life’ study. Journal of Behavior Therapy and Experimental Psychiatry, 50, 257–263. 10.1016/j.jbtep.2015.10.001 [DOI] [PubMed] [Google Scholar]
- Kesting M-L, & Lincoln TM (2013). The relevance of self-esteem and self-schemas to persecutory delusions: A systematic review. Comprehensive Psychiatry, 54(7), 766–789. 10.1016/j.comppsych.2013.03.002 [DOI] [PubMed] [Google Scholar]
- Kramer I, Simons CJP, Wigman JTW, Collip D, Jacobs N, Derom C, Thiery E, van Os J, Myin-Germeys I, & Wichers M (2014). Time-Lagged Moment-to-Moment Interplay Between Negative Affect and Paranoia: New Insights in the Affective Pathway to Psychosis. Schizophrenia Bulletin, 40(2), 278–286. 10.1093/schbul/sbs194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lüdtke T, Hedelt KS, & Westermann S (2023). Predictors of paranoia in the daily lives of people with non-affective psychosis and non-clinical controls: A systematic review of intensive longitudinal studies. Journal of Behavior Therapy and Experimental Psychiatry, 81, 101885. 10.1016/j.jbtep.2023.101885 [DOI] [PubMed] [Google Scholar]
- McGrath JJ, Saha S, Al-Hamzawi A, Alonso J, Bromet EJ, Bruffaerts R, Caldas-de-Almeida JM, Chiu WT, de Jonge P, Fayyad J, Florescu S, Gureje O, Haro JM, Hu C, Kovess-Masfety V, Lepine JP, Lim CCW, Mora MEM, Navarro-Mateu F, … Kessler RC (2015). Psychotic Experiences in the General Population. JAMA Psychiatry, 72(7), 697. 10.1001/jamapsychiatry.2015.0575 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Myin-Germeys I, Klippel A, Steinhart H, & Reininghaus U (2016). Ecological momentary interventions in psychiatry. Current Opinion in Psychiatry, 29(4), 258–263. 10.1097/YCO.0000000000000255 [DOI] [PubMed] [Google Scholar]
- O’Driscoll C, Laing J, & Mason O (2014). Cognitive emotion regulation strategies, alexithymia and dissociation in schizophrenia, a review and meta-analysis. Clinical Psychology Review, 34(6), 482–495. 10.1016/j.cpr.2014.07.002 [DOI] [PubMed] [Google Scholar]
- Rajapakse T, Garcia-Rosales A, Weerawardene S, Cotton S, & Fraser R (2011). Themes of delusions and hallucinations in first-episode psychosis. Early Intervention in Psychiatry, 5(3), 254–258. 10.1111/j.1751-7893.2011.00281.x [DOI] [PubMed] [Google Scholar]
- Smith L, Johns L, & Mitchell R (2017). Characterizing the experience of auditory verbal hallucinations and accompanying delusions in individuals with a diagnosis of bipolar disorder: A systematic review. Bipolar Disorders, 19(6), 417–433. 10.1111/bdi.12520 [DOI] [PubMed] [Google Scholar]
- Thewissen V, Bentall RP, Lecomte T, van Os J, & Myin-Germeys I (2008). Fluctuations in self-esteem and paranoia in the context of daily life. Journal of Abnormal Psychology, 117(1), 143–153. 10.1037/0021-843X.117.1.143 [DOI] [PubMed] [Google Scholar]
- Thewissen V, Bentall RP, Oorschot M, A Campo J, van Lierop T, van Os J, & Myin-Germeys I (2011). Emotions, self-esteem, and paranoid episodes: an experience sampling study. The British Journal of Clinical Psychology / the British Psychological Society, 50(2), 178–195. 10.1348/014466510X508677 [DOI] [PubMed] [Google Scholar]
- Van Os J, & Reininghaus U (2016). Psychosis as a transdiagnostic and extended phenotype in the general population. World Psychiatry, 15(2), 118–124. 10.1002/wps.20310 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang R, Aung MSH, Abdullah S, Brian R, Campbell AT, Choudhury T, Hauser M, Kane J, Merrill M, Scherer EA, Tseng VWS, & Ben-Zeev D (2016). CrossCheck: Towards passive sensing and detection of mental health changes in people with schizophrenia. Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing, 886–897. 10.1145/2971648.2971740 [DOI] [Google Scholar]