Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample Jordan E. DeVylder a, Elizabeth Thompson b, Gloria Reeves c, Jason Schiffman b,n a
School of Social Work, University of Maryland, Baltimore, MD, United States Department of Psychology, University of Maryland, Baltimore County, Baltimore, MD, United States c Division of Child and Adolescent Psychiatry, University of Maryland, Baltimore, MD, United States b
art ic l e i nf o
a b s t r a c t
Article history: Received 6 November 2014 Received in revised form 15 January 2015 Accepted 13 February 2015
Suicide is a leading cause of preventable death. Epidemiological studies have shown strong associations between sub-threshold psychotic experiences and risk for suicidal ideation and behavior. Screens designed to assess psychotic experiences may have clinical utility in improving suicide prevention efforts. In the current study, we hypothesized that the Prodromal Questionnaire-Brief (PQ-B) would reliably distinguish levels of suicidal ideation within a sample of college students (n ¼ 376). As predicted, PQ-B scores varied significantly across levels of suicidal ideation, both when treated as a raw count of sub-threshold psychotic experiences and when taking into account subjective distress associated with those symptoms. In addition, we explored the feasibility of developing a short screen based on the most discriminating items, finding that a six-item version of the PQ-B yielded higher accuracy for detecting elevated suicidal ideation over the full measure. The PQ-B has the potential for clinical utility in detecting groups that might be at increased risk for suicidal ideation. & 2015 Elsevier Ireland Ltd. All rights reserved.
Keywords: Psychotic experiences Suicidal ideation Prodromal Suicidality Psychosis-risk screening Prodromal questionnaire-brief (PQ-B)
1. Introduction Suicide is a leading cause of preventable death, with an estimated 38,000 people dying by suicide each year in the United States alone (Centers for Disease Control and Prevention (CDC), 2010). Intervention efforts, however, have been hindered by difficulty identifying which individuals are at the greatest risk for severe suicidal behavior (Borges et al., 2006). Recent epidemiological studies have shown that sub-threshold psychotic experiences are predictive of suicidal thoughts and attempts among people with ideation, and thus, individuals having psychotic experiences may constitute a group at increased risk (Nishida et al., 2010; Saha et al., 2011; Kelleher et al., 2012, 2013, 2014; Fisher et al., 2013; DeVylder et al., In press; DeVylder and Hilimire, In press). Common in the general population (7.2% estimated lifetime prevalence; Linscott and Van Os, 2013), psychotic experiences qualitatively resemble hallucinations and delusions found in psychotic disorders, but are of insufficient persistence, intensity, or impairment to meet diagnostic criteria. Epidemiological studies provide compelling evidence that psychosis screens may be useful as adjunctive indicators of suicide risk. However, there has not yet
n
Corresponding author. Tel.: þ 1 410 455 1574. E-mail address: [email protected] (J. Schiffman).
been sufficient translational research to draw clinical benefit from these population-level findings, particularly in non-clinical populations. Most notably, epidemiological screens typically do not assess for distress related to psychotic experiences, which is a key component to several psychosis screens. The purpose of this study was to assess the effectiveness of a psychosis-risk screen, the Prodromal Questionnaire-Brief (PQ-B; Loewy et al., 2011), in detecting individuals at elevated risk for suicidal thoughts. We hypothesized that higher scores on the PQ-B would be related to more severe suicidal ideation. In particular, we tested the PQ-B using both its standard scoring method (including distress ratings) and as a raw symptom score, a count of psychotic experiences without considering associated distress. Further, we explored the feasibility of modifying the PQ-B psychosis screen to be tailored towards identification of suicide risk with maximal sensitivity and specificity.
2. Methods 2.1. Participants Participants (N ¼387) were undergraduate students recruited from introductory psychology courses from March 2012 through May 2014 at University of Maryland, Baltimore County (UMBC). Eleven participants (2.8%) were excluded from the study due to
http://dx.doi.org/10.1016/j.psychres.2015.02.007 0165-1781/& 2015 Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
missing demographic data, leaving a final analysis sample of n ¼376. All individuals were over the age of 18 and offered extra credit for their participation in the study. 2.2. Procedure The study was conducted by the Youth FIRST Lab at UMBC and the protocol was approved and overseen by the UMBC Institutional Review Board. Prior to participation, individuals were given an overview of the study after which they read and signed the consent form, keeping a copy for themselves. As part of a larger battery assessing mental health and psychological functioning, all participants completed a demographics form as well as the Prodromal Questionnaire-Brief Version (PQ-B; Loewy et al., 2011), the Beck Depression Inventory-II (BDI-II; Beck et al., 1996), and a treatment history questionnaire developed by the research team that included items assessing current treatment and need for care specific to suicidal thoughts and behavior. Study data were collected via paper and pen self-report measures. Feedback was not provided to participants. Participants were provided referral information for clinical resources in the event of psychological distress or mental health care needs. 2.3. Measures Demographics, including race/ethnicity, age, and sex, were selfreported by participants. Race/ethnicity was recorded into a single categorical variable indicating white, Asian, or other ethnic minority, given the relatively low frequency of many specific racial/ ethnic groups. Substance use was assessed using a measure created by the research team to assess the presence and frequency of alcohol and drug use. Suicidal ideation was assessed using multiple measures. One component of suicide ideation was the suicidality item of the BDIII, which assesses severity of ideation over the past 2 weeks (Beck et al., 1996). Additionally, self-reported current treatment (past 2 months) for suicidal behavior, and/or need for treatment for suicidal behavior were also used as indicators of suicidal ideation. A single ordinal suicide variable was constructed, divided into (1) “lower” suicidal ideation (“I don't have any thoughts of killing myself”) without indicating current treatment or need for care specifically for suicidality; (2) “moderate” suicidal ideation, defined as a score of 1 on the BDI suicide item (“I have thoughts of killing myself, but I would not carry them out”), but without indicating current treatment or need for care specifically for suicidality; and (3) “higher” suicidal ideation, defined as either (a) a score of 1 on the BDI suicide item plus current treatment or need for care specifically for suicidality, or (b) a score of 2 on the BDI suicide item (“I would like to kill myself”). Psychotic experiences were assessed using the PQ-B, a 21-item measure assessing the presence/absence of sub-threshold psychotic experiences, and the distress associated with each symptom rated on a 5-point Likert scale, with higher scores indicating greater distress (Loewy et al., 2011). First published in its longer form in 2005, the PQ-B is a well validated and frequently used measure of psychosis risk (Loewy et al., 2005, 2011; Jarrett et al., 2012; Kline et al., 2012; Addington et al., 2014; Kline and Schiffman, 2014). There are no items on the PQ-B that inquire about suicide or suicidal ideation. PQ-B distress scores were reported as the total sum of distress ratings in accordance with original scoring guidelines. An alternative approach in scoring the PQ-B as a sum of endorsed psychotic experiences (Kline et al., 2014), referred to as PQ-B symptom scores, was also examined. Internal consistency of the PQ-B was excellent in this sample, both with original scoring (i.e. distress scores; a ¼0.90) and when scored as the sum of psychotic experiences (i.e. symptom scores;
a¼ 0.88). The PQ-B was chosen as the measure of psychotic experiences based on prior studies showing strong psychometric properties of this measure within college samples (Kline et al., 2012; Kline and Schiffman, 2014). Depression was measured using the sum score of the BDI-II (Beck et al., 1996), excluding the suicide item. The remaining 20 items were scored each on a four-point scale from 0 to 3, with higher scores indicating greater severity of depressive symptoms. 2.4. Analyses Associations between continuously measured clinical symptoms (PQ-B distress, PQ-B symptom score, and BDI-II) and three levels of suicidality (lower, moderate, and higher) were tested using analysis of variance (ANOVA), with Bonferroni post-hoc tests between pairs of groups. Receiver operating characteristic (ROC) analysis was used to calculate the area under the curve (AUC) and to establish the threshold of maximum sensitivity and specificity for the PQ-B distress, PQ-B symptom score, and BDI-II in distinguishing higher versus moderate suicidal ideation. Logistic regression models were used to calculate odds ratios (OR) to facilitate interpretation of effect sizes of psychotic experiences on suicidality, specifically by showing the increased risk of greater severity of suicidal ideation with each endorsed item on the PQ-B. T-test analyses of individual PQ-B items were used to identify items that indicated higher versus moderate suicidal ideation. A new scale was created using only items (distress scores) that distinguish levels of suicidal ideation at the level of two-tailed alpha ¼0.01 to focus on those items that clearly distinguished levels of suicidal ideation. ROC procedures were repeated for this new exploratory scale, and sensitivity, specificity, positive and negative predictive values, and accuracy were likewise calculated.
3. Results The mean sample age was 20.08 years (S.D. ¼3.29). The sample was half female (n ¼ 191, 50.8%), and of mixed race/ethnicity including white (n ¼134, 35.6%), Asian (n ¼130, 34.6%), and other minority (n ¼112, 29.8%). Substance use was as follows: tobacco: 32.9%; alcohol: 60.5%; marijuana: 29.8%; stimulants: 7.4%; cocaine, amphetamines, sedatives, hallucinogens, or opiates: 20.5%. Substance use was not, however, significantly related to PQ-B scores (no significant Pearson's r between any class of substances, use and frequency of use, with PQ-B symptom or distress scores; data available upon request), and therefore not included in subsequent analyses. Most respondents did not report suicidal thoughts and were included in the “lower” suicidal ideation group (n ¼309, 82.2%). “Moderate” suicidal ideation was reported by 56 individuals (14.9%), and “higher” suicidal ideation was reported by nine respondents (2.4%). Suicide ideation groups did not vary on age, F(d.f.¼2,371) ¼1.41, p ¼0.245, race/ethnicity, χ2(d.f.¼ 2, n ¼374) ¼ 3.65, p¼ 0.455, or sex, χ2(d.f.¼ 2, n¼ 374) ¼1.04, p ¼0.594. Descriptive data for clinical measures are reported in Table 1. All skewness and kurtosis values fall within acceptable range for use of parametric statistics. Table 1 Descriptive data for clinical measures.
PQB-symptom PQB-distress BDI
Range
Mean
S.D.
Skewness
Kurtosis
0–21 0–84 0–41
4.70 13.19 9.89
4.63 15.46 8.16
1.13 1.51 1.22
0.70 1.91 1.55
Note: BDI is scored with the exclusion of the suicidal ideation item.
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
3
Fig. 1. PQ-B scores across levels of suicidal ideation, presented as mean scores with standard errors.
Fig. 2. Receiver Operating Characteristic curves distinguishing higher from moderate risk for suicidal ideation for each PQ-B scoring approach, as well as the BDI.
The PQ-B distress score, F(d.f.¼2,371) ¼39.74, p o0.001, PQ-B symptom score, F(d.f.¼ 2,371) ¼ 32.16, po 0.001, and BDI-II, F(d.f.¼ 2,371) ¼14.98, p o0.001, all varied across levels of suicidality. However, only the PQ-B differentiated between both moderate versus lower (PQ-B distress score: p o0.001; PQ-B symptom score: p o0.001), and between higher versus moderate suicidality (PQ-B distress: p o0.001; PQ-B symptom score: p o0.001) in Bonferroni post-hoc tests (Fig. 1). The mean (S.D.) of PQ-B distress scores was 46.33 (18.79) among those classified as higher suicidal ideation, compared to 21.66 (16.98) for those classified as moderate suicidality and 10.71 (13.38) for those classified as lower suicidal ideation. Suicidality was not significantly related to age, F(d.f.¼ 2,371) ¼1.43, p ¼0.24, gender, χ2(d.f. ¼2, n ¼374) ¼1.09, p ¼0.58, or race/ethnicity, χ2(d.f. ¼4, n ¼374) ¼3.60, p¼ 0.46. ROC curves were used to identify thresholds for maximum sensitivity and specificity for the PQ-B distress scale, PQ-B symptom score, and the BDI (Fig. 2). The maximum sensitivity and specificity of the PQ-B distress scale for distinguishing higher from moderate suicidal ideation were obtained by setting the threshold to scores above 29 (Table 2). The maximum sensitivity and specificity of the PQ-B symptom score were obtained with a threshold of seven endorsed items. The distress score outperformed the symptom score across the entire range of the scale based on the ROC curves.
Each additional psychotic experience endorsed on the PQ-B symptom scale increased the likelihood of any suicidal thoughts/ behavior, OR (95% CI)¼ 1.15 (1.08–1.22), and of higher versus moderate suicidal ideation, OR (95% CI)¼1.28 (1.09–1.51). Each additional point of distress related to psychotic experiences increased likelihood of any suicidal thoughts/behavior, OR (95% CI)¼1.04 (1.03–1.06), and of higher versus moderate suicidal ideation, OR (95% CI)¼1.08 (1.03–1.13). When entered together as independent variables in logistic regression models (without inclusion of the PQ-B symptom score), both the BDI (Wald χ2 ¼14.47, p o0.001) and PQ-B distress (Wald χ2 ¼29.92, po 0.001) distinguished any suicidality (moderate or higher) versus lower (overall model fit: Nagelkerke R2 ¼0.22), but only the PQ-B distress (Wald χ2 ¼ 8.24, p ¼0.004) distinguished higher from moderate suicidal ideation (BDI: Wald χ2 ¼0.02, p ¼0.90; overall model fit: Nagelkerke R2 ¼0.33). To explore potential clinically specific items, we probed individual PQ-B items (using distress scores) that distinguished high versus moderate suicidality. Distress scores were chosen to be consistent with the intended standard scoring procedure for the PQ-B. These items included 1, 2, 9, 14, 15, and 20 (for all t-tests, po 0.01; Table 3). We then performed another ROC analysis with these six items in an exploratory attempt at creating a brief scale to distinguish higher versus moderate suicidal ideation. Results suggest that improved sensitivity, specificity, positive predictive value, and accuracy of the measure for identifying elevated suicidal ideation relative to using the entire measure (Table 2).
4. Discussion This is the first study to our knowledge specifically examining a validated psychosis-risk screener as an indicator of risk for suicidal ideation, building upon a growing epidemiological literature demonstrating that individuals with sub-threshold psychosis are at elevated risk for suicide behavior (Nishida et al., 2010; Saha et al., 2011; Kelleher et al., 2012; 2013; 2014; Fisher et al., 2013; DeVylder and Hilimire, In press). Psychotic experiences tend to occur less frequently than suicidal ideation in the general population, yet appear to co-occur with the more severe range of the suicidal behavior spectrum (i.e., Kelleher et al., 2012), making them prime candidate clinical markers for severity of suicide risk among individuals expressing ideation. We found that high distress scores on the PQ-B were associated with greater levels of suicidal ideation, and that distress scores on the PQ-B appeared to be better statistical indicators of suicidal ideation severity than depression scores (BDI) when compared using logistic regression. Although the number of people in the “higher” suicidal ideation group was very low, results tentatively suggest that the PQ-B was
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
Table 2 Differentiation between higher (n¼9) versus moderate (n¼ 56) suicidal ideation.
Psychosis PQ-B symptom PQ-B distress PQ-B (6)a Depression BDIb a b
AUC
SE
Threshold
Sensitivity
Specificity
PPV
NPV
Accuracy
0.821 0.844 0.949
0.07 0.06 0.03
8.5 29.5 9.5
0.89 0.89 1.00
0.59 0.73 0.82
0.28 0.35 0.41
0.95 0.98 1.00
69.23 75.38 80.00
0.581
0.11
–
–
–
–
–
–
Six-item version of the PQ-B, modified for improved detection of suicidal ideation. Threshold values not calculated for the BDI given that its area under the curve was not significant. BDI is scored with the exclusion of the suicidal ideation item.
Table 3 PQ-B items included in the abridged suicide screen version. Item
Suicide Risk
1. Do familiar surroundings sometimes seem strange, confusing, threatening or unreal to you? 2. Have you heard unusual sounds like banging, clicking, hissing, clapping or ringing in your ears? 9. Do you sometimes get strange feelings on or just beneath your skin, like bugs crawling? 14. Have you been confused at times whether something you experienced was real or imaginary? 15. Do you hold beliefs that other people would find unusual or bizarre? 20. Have you seen things that other people can't see or don't seem to see?
Statistics
Highermean (S.D.)
Moderatemean (S.D.)
2.33 3.22 3.00 3.22 2.67 1.78
0.70 1.18 0.70 1.32 1.04 0.30
(1.87) (1.56) (1.87) (1.48) (1.41) (1.79)
(1.41) (1.64) (1.25) (1.78) (1.31) (1.03)
t63 ¼ 3.08, p ¼0.003 t63 ¼ 3.49, p ¼ 0.001 t63 ¼ 4.77, p o 0.001 t63 ¼ 3.03, p ¼0.004 t63 ¼ 3.44, p ¼ 0.001 t63 ¼ 3.57, p¼ 0.001
Note: Values indicate PQ-B distress scores for each item.
particularly effective in distinguishing “higher” versus “moderate” suicidality. These findings suggest that with additional research, the PQ-B could offer a promising means of identifying individuals who may be at elevated risk of suicidality. The ability to distinguish higher from moderate suicidality was tentatively improved by creating a scale consisting solely of items from the PQ-B that were individually related to suicide severity. Although the development of this brief scale specifically aimed at identifying suicidal ideation is tentatively given our limited sample size for this purpose, it does suggest that this goal might be feasible and worth pursuing in future work with larger nonclinical and possibly clinical samples. Most notably, the abridged six-item PQ-B improved upon the PQ-B's sensitivity and specificity, as well as overall accuracy, suggesting that specific psychosis screen items may be able to provide incrementally useful clinical information regarding suicidal ideation, even in a very brief selfreport assessment. 4.1. Limitations Although suicidal ideation is clinically meaningful irrespective of future outcome (e.g., Chang et al., 2014; Goldney et al., 1991; Reinherz et al., 2006), our measure of suicidal ideation was lacking detail and specificity for particular thoughts and behaviors. Additionally, we created our suicidal ideation groups by combining two different measures that have not been validated to work together for this purpose. However, our measurement had good face validity and our data are in concurrence with past studies. The fact that we found significant relations suggest that associations observed in the present study are robust for measurement variability. Despite evidence suggesting clinically relevant associations between psychosis-risk symptoms and suicidal ideation, due to the cross-sectional design of the current study, causation cannot be inferred from these findings. Future research using large-scale longitudinal designs may serve to establish temporal precedence. Additionally, future studies could use multiple measures including structured assessments of suicide risk (e.g., Columbia-Suicide Severity Rating Scale; Posner et al. (2011)) and psychosis-risk status (e.g., Scale of Psychosis-risk Syndromes; Miller et al. (2003))
to better explore the complexities of suicidality and psychosis proneness. This study also used a convenience sample of college students which may limit generalizability. However, college students are at an age of elevated risk for both suicidal behavior (Kessler et al., 2005) and psychosis (Thompson et al., 2004), making this a population of interest for this research question. The current findings lend support to the use of the PQ-B as a screening tool among this high-risk age group. Finally, although the entire sample was of a reasonable size, “higher” suicidal ideation risk was rare (as is typical in non-help seeking populations), potentially limiting confidence in the reliability of our findings, increasing the likelihood of Type II error, and highlighting the need for replication. Nonetheless, the fact that we found significant differences exploring our a priori hypotheses with such a small subgroup of people at “higher” risk suggest a robustness to the findings. Notably, the effect size of the difference between the higher and moderate risk groups, and the higher and the lower risk groups, on PQ-B distress scores were very large (Cohen's d¼ 1.38 and d ¼2.18 respectively). 4.2. Conclusions Our study is consistent with epidemiological findings that subthreshold psychosis may be a reliable indicator of suicidality (Kelleher et al., 2012; DeVylder et al., In press), and identifies the PQ-B as a particularly useful screening tool within a non-clinical sample for these purposes. Incorporating an indicator of psychosis-risk distress seems to improve the ability to detect suicidal ideation in this sample over simply the presence of psychosis-risk symptoms alone, suggesting the importance of the subjective evaluation of distress associated with symptoms in understanding the link between psychosis-risk and suicidality. Further, our preliminary attempt to abridge the PQ-B to focus specifically on items related to elevated suicidal ideation potentially suggests a feasibility to identifying items within a clinical psychosis-risk screen that can distinguish high from low suicide risk with excellent sensitivity and specificity. This may be useful as a tool to screen for need for services, likely in the context of more comprehensive screening efforts. Using larger samples and
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
following participants over time, future studies should test whether the use of psychosis-risk screens incorporating distress in conjunction with existing suicide screening tools can improve our ability to prospectively predict suicide attempts. Acknowledgment This work was supported in part by funding from the Maryland Department of Health and Mental Hygiene, Mental Hygiene Administration through the Center for Excellence on Early Intervention for Serious Mental Illness (OPASS 14-13717G/M00B4400241). References Addington, J., Stowkowy, J., Weiser, M., 2014. Screening tools for clinical high risk for psychosis. Early Intervention in Psychiatry , http://dx.doi.org/10.1111/eip. 12193. Beck, A.T., Steer, R.A., Brown, G.K., 1996. BDI-II: Beck Depression Inventory Manual. Psychological Corporation, San Antonio, TX. Borges, G., Angst, J., Nock, M.K., Ruscio, A.M., Walters, E.E., Kessler, R.C., 2006. A risk index for 12-month suicide attempts in the National Comorbidity Survey Replication (NCS-R). Psychological Medicine 36 (12), 1747–1757. Centers for Disease Control and Prevention (CDC), 2010. Web-based injury statistics query and reporting system (WISQARS) [Online]. National Center for Injury Prevention and Control, CDC (producer). Retrieved from 〈http://www.cdc.gov. proxy-bc.researchport.umd.edu/ncipc/wisqars/default.htm〉. Chang, W.C., Chen, E.S.M., Hui, C.L.M., Chan, S.K.W., Lee, E.H.M., Chen, E.Y.H., 2014. The relationships of suicidal ideation with symptoms, neurocognitive function, and psychological factors in patients with first-episode psychosis. Schizophrenia Research 157 (1), 12–18. DeVylder, J.E., Hilimire, M.R., 2015. Suicide risk, stress sensitivity, and self-esteem among young adults reporting auditory hallucinations. Health & Social Work, In press. DeVylder, J.E., Lukens, E.P., Link, B.G., Lieberman, J.A., 2015. Suicidal ideation and suicide attempts among adults with psychotic experiences: data from the Collaborative Psychiatric Epidemiology Surveys. JAMA Psychiatry, http://dx.doi. org/10.1001/jamapsychiatry.2014.2663. Fisher, H.L., Caspi, A., Poulton, R., Meier, M.H., Houts, R., Harrington, H., Arseneault, L., Moffitt, T.E., 2013. Specificity of childhood psychotic symptoms for predicting schizophrenia by 38 years of age: a birth cohort study. Psychological Medicine 43 (10), 2077–2086. Goldney, R.D., Smith, S., Winefield, A.H., Tiggeman, M., Winefield, H.R., 1991. Suicidal ideation: its enduring nature and associated morbidity. Acta Psychiatrica Scandinavica 83 (2), 115–120. Jarrett, M., Craig, T., Parrott, J., Forrester, A., Winton-Brown, T., Maguire, H., MaGuire, P., Valmaggia, L., 2012. Identifying men at ultra high risk of psychosis in a prison population. Schizophrenia Research 136 (1), 1–6. Kelleher, I., Corcoran, P., Keeley, H., Wigman, J.T., Devlin, N., Ramsay, H., Wasserman, C., Carli, V., Sarchiapone, M., Hoven, C., Wasserman, D., Cannon, M., 2013. Psychotic symptoms and population risk for suicide attempt: a prospective cohort study. JAMA Psychiatry 70 (9), 940–948.
5
Kelleher, I., Devlin, N., Wigman, J.T., Kehoe, A., Murtagh, A., Fitzpatrick, C., Cannon, M., 2014. Psychotic experiences in a mental health clinic sample: implications for suicidality, multimorbidity and functioning. Psychological Medicine 44 (08), 1615–1624. Kelleher, I., Lynch, F., Harley, M., Molloy, C., Roddy, S., Fitzpatrick, C., Cannon, M., 2012. Psychotic symptoms in adolescence index risk for suicidal behavior: findings from 2 population-based case-control clinical interview studies. Archives of General Psychiatry 69 (12), 1277–1283. Kessler, R.C., Berglund, P., Borges, G., Nock, M., Wang, P.S., 2005. Trends in suicide ideation, plans, gestures, and attempts in the United States, 1990–1992 to 2001–2003. JAMA 293 (20), 2487–2495. Kline, E., Schiffman, J., 2014. Psychosis risk screening: a systematic review. Schizophrenia Research 158 (1), 11–18. Kline, E., Thompson, E., Bussell, K., Pitts, S.C., Reeves, G., Schiffman, J., 2014. Psychosis-like experiences and distress among adolescents using mental health services. Schizophrenia Research 152 (2), 498–502. Kline, E., Wilson, C., Ereshefsky, S., Tsuji, T., Schiffman, J., Pitts, S., Reeves, G., 2012. Convergent and discriminant validity of attenuated psychosis screening tools. Schizophrenia Research 134 (1), 49–53. Linscott, R.J., Van Os, J., 2013. An updated and conservative systematic review and meta-analysis of epidemiological evidence on psychotic experiences in children and adults: on the pathway from proneness to persistence to dimensional expression across mental disorders. Psychological Medicine 43 (06), 1133–1149. Loewy, R.L., Bearden, C.E., Johnson, J.K., Raine, A., Cannon, T.D., 2005. The prodromal questionnaire (PQ): preliminary validation of a self-report screening measure for prodromal and psychotic syndromes. Schizophrenia Research 79 (1), 117–125. Loewy, R.L., Pearson, R., Vinogradov, S., Bearden, C.E., Cannon, T.D., 2011. Psychosis risk screening with the Prodromal Questionnaire-Brief version (PQ-B). Schizophrenia Research 129 (1), 42–46. Miller, T.J., McGlashan, T.H., Rosen, J.L., Cadenhead, K., Ventura, J., McFarlane, W., Perkins, D.O., Pearlson, G.D., Woods, S.W., 2003. Prodromal assessment with the structured interview for prodromal syndromes and the scale of prodromal symptoms: predictive validity, interrater reliability, and training to reliability. Schizophrenia Bulletin 29 (4), 703. Nishida, A., Sasaki, T., Nishimura, Y., Tanii, H., Hara, N., Inoue, K., Yamada, T., Takami, T., Shimodera, S., Itokawa, M., Asukai, N., Okazaki, Y., 2010. Psychotic‐like experiences are associated with suicidal feelings and deliberate self‐harm behaviors in adolescents aged 12–15 years. Acta Psychiatrica Scandinavica 121 (4), 301–307. Posner, K., Brown, G.K., Stanley, B., Brent, D.A., Yershova, K.V., Oquendo, M.A., Currier, G.W., Melvin, G.A., Greenhill, L., Shen, S., Mann, J.J., 2011. The Columbiasuicide severity rating scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults. American Journal of Psychiatry 168 (12), 1266–1277. Reinherz, H., Tanner, J., Berger, S., Beardslee, W., Fitzmaurice, G., 2006. Adolescent suicidal ideation as predictive of psychopathology, suicidal behavior, and compromised functioning at age 30. American Journal of Psychiatry 163 (7), 1226–1232. Saha, S., Scott, J.G., Johnston, A.K., Slade, T.N., Varghese, D., Carter, G.L., McGrath, J.J., 2011. The association between delusional-like experiences and suicidal thoughts and behaviour. Schizophrenia Research 132 (2), 197–202. Thompson, J.L., Pogue-Geile, M.F., Grace, A.A., 2004. Developmental pathology, dopamine, and stress: a model for the age of onset of schizophrenia symptoms. Schizophrenia Bulletin 30 (4), 875–900.
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
Contents lists available at ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample Jordan E. DeVylder a, Elizabeth Thompson b, Gloria Reeves c, Jason Schiffman b,n a
School of Social Work, University of Maryland, Baltimore, MD, United States Department of Psychology, University of Maryland, Baltimore County, Baltimore, MD, United States c Division of Child and Adolescent Psychiatry, University of Maryland, Baltimore, MD, United States b
art ic l e i nf o
a b s t r a c t
Article history: Received 6 November 2014 Received in revised form 15 January 2015 Accepted 13 February 2015
Suicide is a leading cause of preventable death. Epidemiological studies have shown strong associations between sub-threshold psychotic experiences and risk for suicidal ideation and behavior. Screens designed to assess psychotic experiences may have clinical utility in improving suicide prevention efforts. In the current study, we hypothesized that the Prodromal Questionnaire-Brief (PQ-B) would reliably distinguish levels of suicidal ideation within a sample of college students (n ¼ 376). As predicted, PQ-B scores varied significantly across levels of suicidal ideation, both when treated as a raw count of sub-threshold psychotic experiences and when taking into account subjective distress associated with those symptoms. In addition, we explored the feasibility of developing a short screen based on the most discriminating items, finding that a six-item version of the PQ-B yielded higher accuracy for detecting elevated suicidal ideation over the full measure. The PQ-B has the potential for clinical utility in detecting groups that might be at increased risk for suicidal ideation. & 2015 Elsevier Ireland Ltd. All rights reserved.
Keywords: Psychotic experiences Suicidal ideation Prodromal Suicidality Psychosis-risk screening Prodromal questionnaire-brief (PQ-B)
1. Introduction Suicide is a leading cause of preventable death, with an estimated 38,000 people dying by suicide each year in the United States alone (Centers for Disease Control and Prevention (CDC), 2010). Intervention efforts, however, have been hindered by difficulty identifying which individuals are at the greatest risk for severe suicidal behavior (Borges et al., 2006). Recent epidemiological studies have shown that sub-threshold psychotic experiences are predictive of suicidal thoughts and attempts among people with ideation, and thus, individuals having psychotic experiences may constitute a group at increased risk (Nishida et al., 2010; Saha et al., 2011; Kelleher et al., 2012, 2013, 2014; Fisher et al., 2013; DeVylder et al., In press; DeVylder and Hilimire, In press). Common in the general population (7.2% estimated lifetime prevalence; Linscott and Van Os, 2013), psychotic experiences qualitatively resemble hallucinations and delusions found in psychotic disorders, but are of insufficient persistence, intensity, or impairment to meet diagnostic criteria. Epidemiological studies provide compelling evidence that psychosis screens may be useful as adjunctive indicators of suicide risk. However, there has not yet
n
Corresponding author. Tel.: þ 1 410 455 1574. E-mail address: [email protected] (J. Schiffman).
been sufficient translational research to draw clinical benefit from these population-level findings, particularly in non-clinical populations. Most notably, epidemiological screens typically do not assess for distress related to psychotic experiences, which is a key component to several psychosis screens. The purpose of this study was to assess the effectiveness of a psychosis-risk screen, the Prodromal Questionnaire-Brief (PQ-B; Loewy et al., 2011), in detecting individuals at elevated risk for suicidal thoughts. We hypothesized that higher scores on the PQ-B would be related to more severe suicidal ideation. In particular, we tested the PQ-B using both its standard scoring method (including distress ratings) and as a raw symptom score, a count of psychotic experiences without considering associated distress. Further, we explored the feasibility of modifying the PQ-B psychosis screen to be tailored towards identification of suicide risk with maximal sensitivity and specificity.
2. Methods 2.1. Participants Participants (N ¼387) were undergraduate students recruited from introductory psychology courses from March 2012 through May 2014 at University of Maryland, Baltimore County (UMBC). Eleven participants (2.8%) were excluded from the study due to
http://dx.doi.org/10.1016/j.psychres.2015.02.007 0165-1781/& 2015 Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
missing demographic data, leaving a final analysis sample of n ¼376. All individuals were over the age of 18 and offered extra credit for their participation in the study. 2.2. Procedure The study was conducted by the Youth FIRST Lab at UMBC and the protocol was approved and overseen by the UMBC Institutional Review Board. Prior to participation, individuals were given an overview of the study after which they read and signed the consent form, keeping a copy for themselves. As part of a larger battery assessing mental health and psychological functioning, all participants completed a demographics form as well as the Prodromal Questionnaire-Brief Version (PQ-B; Loewy et al., 2011), the Beck Depression Inventory-II (BDI-II; Beck et al., 1996), and a treatment history questionnaire developed by the research team that included items assessing current treatment and need for care specific to suicidal thoughts and behavior. Study data were collected via paper and pen self-report measures. Feedback was not provided to participants. Participants were provided referral information for clinical resources in the event of psychological distress or mental health care needs. 2.3. Measures Demographics, including race/ethnicity, age, and sex, were selfreported by participants. Race/ethnicity was recorded into a single categorical variable indicating white, Asian, or other ethnic minority, given the relatively low frequency of many specific racial/ ethnic groups. Substance use was assessed using a measure created by the research team to assess the presence and frequency of alcohol and drug use. Suicidal ideation was assessed using multiple measures. One component of suicide ideation was the suicidality item of the BDIII, which assesses severity of ideation over the past 2 weeks (Beck et al., 1996). Additionally, self-reported current treatment (past 2 months) for suicidal behavior, and/or need for treatment for suicidal behavior were also used as indicators of suicidal ideation. A single ordinal suicide variable was constructed, divided into (1) “lower” suicidal ideation (“I don't have any thoughts of killing myself”) without indicating current treatment or need for care specifically for suicidality; (2) “moderate” suicidal ideation, defined as a score of 1 on the BDI suicide item (“I have thoughts of killing myself, but I would not carry them out”), but without indicating current treatment or need for care specifically for suicidality; and (3) “higher” suicidal ideation, defined as either (a) a score of 1 on the BDI suicide item plus current treatment or need for care specifically for suicidality, or (b) a score of 2 on the BDI suicide item (“I would like to kill myself”). Psychotic experiences were assessed using the PQ-B, a 21-item measure assessing the presence/absence of sub-threshold psychotic experiences, and the distress associated with each symptom rated on a 5-point Likert scale, with higher scores indicating greater distress (Loewy et al., 2011). First published in its longer form in 2005, the PQ-B is a well validated and frequently used measure of psychosis risk (Loewy et al., 2005, 2011; Jarrett et al., 2012; Kline et al., 2012; Addington et al., 2014; Kline and Schiffman, 2014). There are no items on the PQ-B that inquire about suicide or suicidal ideation. PQ-B distress scores were reported as the total sum of distress ratings in accordance with original scoring guidelines. An alternative approach in scoring the PQ-B as a sum of endorsed psychotic experiences (Kline et al., 2014), referred to as PQ-B symptom scores, was also examined. Internal consistency of the PQ-B was excellent in this sample, both with original scoring (i.e. distress scores; a ¼0.90) and when scored as the sum of psychotic experiences (i.e. symptom scores;
a¼ 0.88). The PQ-B was chosen as the measure of psychotic experiences based on prior studies showing strong psychometric properties of this measure within college samples (Kline et al., 2012; Kline and Schiffman, 2014). Depression was measured using the sum score of the BDI-II (Beck et al., 1996), excluding the suicide item. The remaining 20 items were scored each on a four-point scale from 0 to 3, with higher scores indicating greater severity of depressive symptoms. 2.4. Analyses Associations between continuously measured clinical symptoms (PQ-B distress, PQ-B symptom score, and BDI-II) and three levels of suicidality (lower, moderate, and higher) were tested using analysis of variance (ANOVA), with Bonferroni post-hoc tests between pairs of groups. Receiver operating characteristic (ROC) analysis was used to calculate the area under the curve (AUC) and to establish the threshold of maximum sensitivity and specificity for the PQ-B distress, PQ-B symptom score, and BDI-II in distinguishing higher versus moderate suicidal ideation. Logistic regression models were used to calculate odds ratios (OR) to facilitate interpretation of effect sizes of psychotic experiences on suicidality, specifically by showing the increased risk of greater severity of suicidal ideation with each endorsed item on the PQ-B. T-test analyses of individual PQ-B items were used to identify items that indicated higher versus moderate suicidal ideation. A new scale was created using only items (distress scores) that distinguish levels of suicidal ideation at the level of two-tailed alpha ¼0.01 to focus on those items that clearly distinguished levels of suicidal ideation. ROC procedures were repeated for this new exploratory scale, and sensitivity, specificity, positive and negative predictive values, and accuracy were likewise calculated.
3. Results The mean sample age was 20.08 years (S.D. ¼3.29). The sample was half female (n ¼ 191, 50.8%), and of mixed race/ethnicity including white (n ¼134, 35.6%), Asian (n ¼130, 34.6%), and other minority (n ¼112, 29.8%). Substance use was as follows: tobacco: 32.9%; alcohol: 60.5%; marijuana: 29.8%; stimulants: 7.4%; cocaine, amphetamines, sedatives, hallucinogens, or opiates: 20.5%. Substance use was not, however, significantly related to PQ-B scores (no significant Pearson's r between any class of substances, use and frequency of use, with PQ-B symptom or distress scores; data available upon request), and therefore not included in subsequent analyses. Most respondents did not report suicidal thoughts and were included in the “lower” suicidal ideation group (n ¼309, 82.2%). “Moderate” suicidal ideation was reported by 56 individuals (14.9%), and “higher” suicidal ideation was reported by nine respondents (2.4%). Suicide ideation groups did not vary on age, F(d.f.¼2,371) ¼1.41, p ¼0.245, race/ethnicity, χ2(d.f.¼ 2, n ¼374) ¼ 3.65, p¼ 0.455, or sex, χ2(d.f.¼ 2, n¼ 374) ¼1.04, p ¼0.594. Descriptive data for clinical measures are reported in Table 1. All skewness and kurtosis values fall within acceptable range for use of parametric statistics. Table 1 Descriptive data for clinical measures.
PQB-symptom PQB-distress BDI
Range
Mean
S.D.
Skewness
Kurtosis
0–21 0–84 0–41
4.70 13.19 9.89
4.63 15.46 8.16
1.13 1.51 1.22
0.70 1.91 1.55
Note: BDI is scored with the exclusion of the suicidal ideation item.
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
3
Fig. 1. PQ-B scores across levels of suicidal ideation, presented as mean scores with standard errors.
Fig. 2. Receiver Operating Characteristic curves distinguishing higher from moderate risk for suicidal ideation for each PQ-B scoring approach, as well as the BDI.
The PQ-B distress score, F(d.f.¼2,371) ¼39.74, p o0.001, PQ-B symptom score, F(d.f.¼ 2,371) ¼ 32.16, po 0.001, and BDI-II, F(d.f.¼ 2,371) ¼14.98, p o0.001, all varied across levels of suicidality. However, only the PQ-B differentiated between both moderate versus lower (PQ-B distress score: p o0.001; PQ-B symptom score: p o0.001), and between higher versus moderate suicidality (PQ-B distress: p o0.001; PQ-B symptom score: p o0.001) in Bonferroni post-hoc tests (Fig. 1). The mean (S.D.) of PQ-B distress scores was 46.33 (18.79) among those classified as higher suicidal ideation, compared to 21.66 (16.98) for those classified as moderate suicidality and 10.71 (13.38) for those classified as lower suicidal ideation. Suicidality was not significantly related to age, F(d.f.¼ 2,371) ¼1.43, p ¼0.24, gender, χ2(d.f. ¼2, n ¼374) ¼1.09, p ¼0.58, or race/ethnicity, χ2(d.f. ¼4, n ¼374) ¼3.60, p¼ 0.46. ROC curves were used to identify thresholds for maximum sensitivity and specificity for the PQ-B distress scale, PQ-B symptom score, and the BDI (Fig. 2). The maximum sensitivity and specificity of the PQ-B distress scale for distinguishing higher from moderate suicidal ideation were obtained by setting the threshold to scores above 29 (Table 2). The maximum sensitivity and specificity of the PQ-B symptom score were obtained with a threshold of seven endorsed items. The distress score outperformed the symptom score across the entire range of the scale based on the ROC curves.
Each additional psychotic experience endorsed on the PQ-B symptom scale increased the likelihood of any suicidal thoughts/ behavior, OR (95% CI)¼ 1.15 (1.08–1.22), and of higher versus moderate suicidal ideation, OR (95% CI)¼1.28 (1.09–1.51). Each additional point of distress related to psychotic experiences increased likelihood of any suicidal thoughts/behavior, OR (95% CI)¼1.04 (1.03–1.06), and of higher versus moderate suicidal ideation, OR (95% CI)¼1.08 (1.03–1.13). When entered together as independent variables in logistic regression models (without inclusion of the PQ-B symptom score), both the BDI (Wald χ2 ¼14.47, p o0.001) and PQ-B distress (Wald χ2 ¼29.92, po 0.001) distinguished any suicidality (moderate or higher) versus lower (overall model fit: Nagelkerke R2 ¼0.22), but only the PQ-B distress (Wald χ2 ¼ 8.24, p ¼0.004) distinguished higher from moderate suicidal ideation (BDI: Wald χ2 ¼0.02, p ¼0.90; overall model fit: Nagelkerke R2 ¼0.33). To explore potential clinically specific items, we probed individual PQ-B items (using distress scores) that distinguished high versus moderate suicidality. Distress scores were chosen to be consistent with the intended standard scoring procedure for the PQ-B. These items included 1, 2, 9, 14, 15, and 20 (for all t-tests, po 0.01; Table 3). We then performed another ROC analysis with these six items in an exploratory attempt at creating a brief scale to distinguish higher versus moderate suicidal ideation. Results suggest that improved sensitivity, specificity, positive predictive value, and accuracy of the measure for identifying elevated suicidal ideation relative to using the entire measure (Table 2).
4. Discussion This is the first study to our knowledge specifically examining a validated psychosis-risk screener as an indicator of risk for suicidal ideation, building upon a growing epidemiological literature demonstrating that individuals with sub-threshold psychosis are at elevated risk for suicide behavior (Nishida et al., 2010; Saha et al., 2011; Kelleher et al., 2012; 2013; 2014; Fisher et al., 2013; DeVylder and Hilimire, In press). Psychotic experiences tend to occur less frequently than suicidal ideation in the general population, yet appear to co-occur with the more severe range of the suicidal behavior spectrum (i.e., Kelleher et al., 2012), making them prime candidate clinical markers for severity of suicide risk among individuals expressing ideation. We found that high distress scores on the PQ-B were associated with greater levels of suicidal ideation, and that distress scores on the PQ-B appeared to be better statistical indicators of suicidal ideation severity than depression scores (BDI) when compared using logistic regression. Although the number of people in the “higher” suicidal ideation group was very low, results tentatively suggest that the PQ-B was
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
Table 2 Differentiation between higher (n¼9) versus moderate (n¼ 56) suicidal ideation.
Psychosis PQ-B symptom PQ-B distress PQ-B (6)a Depression BDIb a b
AUC
SE
Threshold
Sensitivity
Specificity
PPV
NPV
Accuracy
0.821 0.844 0.949
0.07 0.06 0.03
8.5 29.5 9.5
0.89 0.89 1.00
0.59 0.73 0.82
0.28 0.35 0.41
0.95 0.98 1.00
69.23 75.38 80.00
0.581
0.11
–
–
–
–
–
–
Six-item version of the PQ-B, modified for improved detection of suicidal ideation. Threshold values not calculated for the BDI given that its area under the curve was not significant. BDI is scored with the exclusion of the suicidal ideation item.
Table 3 PQ-B items included in the abridged suicide screen version. Item
Suicide Risk
1. Do familiar surroundings sometimes seem strange, confusing, threatening or unreal to you? 2. Have you heard unusual sounds like banging, clicking, hissing, clapping or ringing in your ears? 9. Do you sometimes get strange feelings on or just beneath your skin, like bugs crawling? 14. Have you been confused at times whether something you experienced was real or imaginary? 15. Do you hold beliefs that other people would find unusual or bizarre? 20. Have you seen things that other people can't see or don't seem to see?
Statistics
Highermean (S.D.)
Moderatemean (S.D.)
2.33 3.22 3.00 3.22 2.67 1.78
0.70 1.18 0.70 1.32 1.04 0.30
(1.87) (1.56) (1.87) (1.48) (1.41) (1.79)
(1.41) (1.64) (1.25) (1.78) (1.31) (1.03)
t63 ¼ 3.08, p ¼0.003 t63 ¼ 3.49, p ¼ 0.001 t63 ¼ 4.77, p o 0.001 t63 ¼ 3.03, p ¼0.004 t63 ¼ 3.44, p ¼ 0.001 t63 ¼ 3.57, p¼ 0.001
Note: Values indicate PQ-B distress scores for each item.
particularly effective in distinguishing “higher” versus “moderate” suicidality. These findings suggest that with additional research, the PQ-B could offer a promising means of identifying individuals who may be at elevated risk of suicidality. The ability to distinguish higher from moderate suicidality was tentatively improved by creating a scale consisting solely of items from the PQ-B that were individually related to suicide severity. Although the development of this brief scale specifically aimed at identifying suicidal ideation is tentatively given our limited sample size for this purpose, it does suggest that this goal might be feasible and worth pursuing in future work with larger nonclinical and possibly clinical samples. Most notably, the abridged six-item PQ-B improved upon the PQ-B's sensitivity and specificity, as well as overall accuracy, suggesting that specific psychosis screen items may be able to provide incrementally useful clinical information regarding suicidal ideation, even in a very brief selfreport assessment. 4.1. Limitations Although suicidal ideation is clinically meaningful irrespective of future outcome (e.g., Chang et al., 2014; Goldney et al., 1991; Reinherz et al., 2006), our measure of suicidal ideation was lacking detail and specificity for particular thoughts and behaviors. Additionally, we created our suicidal ideation groups by combining two different measures that have not been validated to work together for this purpose. However, our measurement had good face validity and our data are in concurrence with past studies. The fact that we found significant relations suggest that associations observed in the present study are robust for measurement variability. Despite evidence suggesting clinically relevant associations between psychosis-risk symptoms and suicidal ideation, due to the cross-sectional design of the current study, causation cannot be inferred from these findings. Future research using large-scale longitudinal designs may serve to establish temporal precedence. Additionally, future studies could use multiple measures including structured assessments of suicide risk (e.g., Columbia-Suicide Severity Rating Scale; Posner et al. (2011)) and psychosis-risk status (e.g., Scale of Psychosis-risk Syndromes; Miller et al. (2003))
to better explore the complexities of suicidality and psychosis proneness. This study also used a convenience sample of college students which may limit generalizability. However, college students are at an age of elevated risk for both suicidal behavior (Kessler et al., 2005) and psychosis (Thompson et al., 2004), making this a population of interest for this research question. The current findings lend support to the use of the PQ-B as a screening tool among this high-risk age group. Finally, although the entire sample was of a reasonable size, “higher” suicidal ideation risk was rare (as is typical in non-help seeking populations), potentially limiting confidence in the reliability of our findings, increasing the likelihood of Type II error, and highlighting the need for replication. Nonetheless, the fact that we found significant differences exploring our a priori hypotheses with such a small subgroup of people at “higher” risk suggest a robustness to the findings. Notably, the effect size of the difference between the higher and moderate risk groups, and the higher and the lower risk groups, on PQ-B distress scores were very large (Cohen's d¼ 1.38 and d ¼2.18 respectively). 4.2. Conclusions Our study is consistent with epidemiological findings that subthreshold psychosis may be a reliable indicator of suicidality (Kelleher et al., 2012; DeVylder et al., In press), and identifies the PQ-B as a particularly useful screening tool within a non-clinical sample for these purposes. Incorporating an indicator of psychosis-risk distress seems to improve the ability to detect suicidal ideation in this sample over simply the presence of psychosis-risk symptoms alone, suggesting the importance of the subjective evaluation of distress associated with symptoms in understanding the link between psychosis-risk and suicidality. Further, our preliminary attempt to abridge the PQ-B to focus specifically on items related to elevated suicidal ideation potentially suggests a feasibility to identifying items within a clinical psychosis-risk screen that can distinguish high from low suicide risk with excellent sensitivity and specificity. This may be useful as a tool to screen for need for services, likely in the context of more comprehensive screening efforts. Using larger samples and
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
J.E. DeVylder et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
following participants over time, future studies should test whether the use of psychosis-risk screens incorporating distress in conjunction with existing suicide screening tools can improve our ability to prospectively predict suicide attempts. Acknowledgment This work was supported in part by funding from the Maryland Department of Health and Mental Hygiene, Mental Hygiene Administration through the Center for Excellence on Early Intervention for Serious Mental Illness (OPASS 14-13717G/M00B4400241). References Addington, J., Stowkowy, J., Weiser, M., 2014. Screening tools for clinical high risk for psychosis. Early Intervention in Psychiatry , http://dx.doi.org/10.1111/eip. 12193. Beck, A.T., Steer, R.A., Brown, G.K., 1996. BDI-II: Beck Depression Inventory Manual. Psychological Corporation, San Antonio, TX. Borges, G., Angst, J., Nock, M.K., Ruscio, A.M., Walters, E.E., Kessler, R.C., 2006. A risk index for 12-month suicide attempts in the National Comorbidity Survey Replication (NCS-R). Psychological Medicine 36 (12), 1747–1757. Centers for Disease Control and Prevention (CDC), 2010. Web-based injury statistics query and reporting system (WISQARS) [Online]. National Center for Injury Prevention and Control, CDC (producer). Retrieved from 〈http://www.cdc.gov. proxy-bc.researchport.umd.edu/ncipc/wisqars/default.htm〉. Chang, W.C., Chen, E.S.M., Hui, C.L.M., Chan, S.K.W., Lee, E.H.M., Chen, E.Y.H., 2014. The relationships of suicidal ideation with symptoms, neurocognitive function, and psychological factors in patients with first-episode psychosis. Schizophrenia Research 157 (1), 12–18. DeVylder, J.E., Hilimire, M.R., 2015. Suicide risk, stress sensitivity, and self-esteem among young adults reporting auditory hallucinations. Health & Social Work, In press. DeVylder, J.E., Lukens, E.P., Link, B.G., Lieberman, J.A., 2015. Suicidal ideation and suicide attempts among adults with psychotic experiences: data from the Collaborative Psychiatric Epidemiology Surveys. JAMA Psychiatry, http://dx.doi. org/10.1001/jamapsychiatry.2014.2663. Fisher, H.L., Caspi, A., Poulton, R., Meier, M.H., Houts, R., Harrington, H., Arseneault, L., Moffitt, T.E., 2013. Specificity of childhood psychotic symptoms for predicting schizophrenia by 38 years of age: a birth cohort study. Psychological Medicine 43 (10), 2077–2086. Goldney, R.D., Smith, S., Winefield, A.H., Tiggeman, M., Winefield, H.R., 1991. Suicidal ideation: its enduring nature and associated morbidity. Acta Psychiatrica Scandinavica 83 (2), 115–120. Jarrett, M., Craig, T., Parrott, J., Forrester, A., Winton-Brown, T., Maguire, H., MaGuire, P., Valmaggia, L., 2012. Identifying men at ultra high risk of psychosis in a prison population. Schizophrenia Research 136 (1), 1–6. Kelleher, I., Corcoran, P., Keeley, H., Wigman, J.T., Devlin, N., Ramsay, H., Wasserman, C., Carli, V., Sarchiapone, M., Hoven, C., Wasserman, D., Cannon, M., 2013. Psychotic symptoms and population risk for suicide attempt: a prospective cohort study. JAMA Psychiatry 70 (9), 940–948.
5
Kelleher, I., Devlin, N., Wigman, J.T., Kehoe, A., Murtagh, A., Fitzpatrick, C., Cannon, M., 2014. Psychotic experiences in a mental health clinic sample: implications for suicidality, multimorbidity and functioning. Psychological Medicine 44 (08), 1615–1624. Kelleher, I., Lynch, F., Harley, M., Molloy, C., Roddy, S., Fitzpatrick, C., Cannon, M., 2012. Psychotic symptoms in adolescence index risk for suicidal behavior: findings from 2 population-based case-control clinical interview studies. Archives of General Psychiatry 69 (12), 1277–1283. Kessler, R.C., Berglund, P., Borges, G., Nock, M., Wang, P.S., 2005. Trends in suicide ideation, plans, gestures, and attempts in the United States, 1990–1992 to 2001–2003. JAMA 293 (20), 2487–2495. Kline, E., Schiffman, J., 2014. Psychosis risk screening: a systematic review. Schizophrenia Research 158 (1), 11–18. Kline, E., Thompson, E., Bussell, K., Pitts, S.C., Reeves, G., Schiffman, J., 2014. Psychosis-like experiences and distress among adolescents using mental health services. Schizophrenia Research 152 (2), 498–502. Kline, E., Wilson, C., Ereshefsky, S., Tsuji, T., Schiffman, J., Pitts, S., Reeves, G., 2012. Convergent and discriminant validity of attenuated psychosis screening tools. Schizophrenia Research 134 (1), 49–53. Linscott, R.J., Van Os, J., 2013. An updated and conservative systematic review and meta-analysis of epidemiological evidence on psychotic experiences in children and adults: on the pathway from proneness to persistence to dimensional expression across mental disorders. Psychological Medicine 43 (06), 1133–1149. Loewy, R.L., Bearden, C.E., Johnson, J.K., Raine, A., Cannon, T.D., 2005. The prodromal questionnaire (PQ): preliminary validation of a self-report screening measure for prodromal and psychotic syndromes. Schizophrenia Research 79 (1), 117–125. Loewy, R.L., Pearson, R., Vinogradov, S., Bearden, C.E., Cannon, T.D., 2011. Psychosis risk screening with the Prodromal Questionnaire-Brief version (PQ-B). Schizophrenia Research 129 (1), 42–46. Miller, T.J., McGlashan, T.H., Rosen, J.L., Cadenhead, K., Ventura, J., McFarlane, W., Perkins, D.O., Pearlson, G.D., Woods, S.W., 2003. Prodromal assessment with the structured interview for prodromal syndromes and the scale of prodromal symptoms: predictive validity, interrater reliability, and training to reliability. Schizophrenia Bulletin 29 (4), 703. Nishida, A., Sasaki, T., Nishimura, Y., Tanii, H., Hara, N., Inoue, K., Yamada, T., Takami, T., Shimodera, S., Itokawa, M., Asukai, N., Okazaki, Y., 2010. Psychotic‐like experiences are associated with suicidal feelings and deliberate self‐harm behaviors in adolescents aged 12–15 years. Acta Psychiatrica Scandinavica 121 (4), 301–307. Posner, K., Brown, G.K., Stanley, B., Brent, D.A., Yershova, K.V., Oquendo, M.A., Currier, G.W., Melvin, G.A., Greenhill, L., Shen, S., Mann, J.J., 2011. The Columbiasuicide severity rating scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults. American Journal of Psychiatry 168 (12), 1266–1277. Reinherz, H., Tanner, J., Berger, S., Beardslee, W., Fitzmaurice, G., 2006. Adolescent suicidal ideation as predictive of psychopathology, suicidal behavior, and compromised functioning at age 30. American Journal of Psychiatry 163 (7), 1226–1232. Saha, S., Scott, J.G., Johnston, A.K., Slade, T.N., Varghese, D., Carter, G.L., McGrath, J.J., 2011. The association between delusional-like experiences and suicidal thoughts and behaviour. Schizophrenia Research 132 (2), 197–202. Thompson, J.L., Pogue-Geile, M.F., Grace, A.A., 2004. Developmental pathology, dopamine, and stress: a model for the age of onset of schizophrenia symptoms. Schizophrenia Bulletin 30 (4), 875–900.
Please cite this article as: DeVylder, J.E., et al., Psychotic experiences as indicators of suicidal ideation in a non-clinical college sample. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.02.007i
