Early Expression of Negative/Disorganized Symptoms Predicting Psychotic Experiences and Subsequent Clinical Psychosis: A 10-Year Study
Abstract
Objective:
The cognitive and motivational impairments observed in psychotic disorders may reflect early developmental alterations that, when combined with later environmental exposures, may drive the onset of positive psychotic symptoms. The epidemiological predictions of this model were tested.
Method:
A longitudinal prospective cohort study (the Early Developmental Stages of Psychopathology Study) was conducted with a representative general population sample of adolescents and young adults from Munich (N=3,021), who were 14–24 years of age at baseline. Sociodemographic factors, environmental exposures, and measures of psychopathology and associated clinical relevance were assessed across three waves, covering a period of up to 10 years, by clinical psychologists using the Composite International Diagnostic Interview.
Results:
Both negative/disorganized and positive psychotic symptoms were frequent (5-year cumulative prevalence rates of around 12%) and occurred in combination more often than predicted by chance. Negative/disorganized symptoms revealed a pattern of sociodemographic associations indicative of developmental impairment, whereas the positive symptoms were associated with environmental exposures such as trauma, cannabis use, and urbanicity. Negative/disorganized symptoms predicted positive symptoms over time, and co-occurrence of positive and negative/disorganized symptoms was predictive of clinical relevance in terms of secondary functional impairment and help-seeking behavior.
Conclusion:
The results suggest that the negative/disorganized features associated with psychotic disorder are distributed at the population level and drive the ontogenesis of positive psychotic experiences after exposure to environmental risks, increasing the likelihood of impairment and need for care.
Systematic review of general population surveys indicates that the experiences associated with psychotic disorder, such as paranoid delusional thinking and auditory hallucinations, are observed—in an attenuated form—in 5%–8% of respondents (1, 2). These attenuated expressions may be conceived as the behavioral expression of an underlying distributed liability for psychosis (3). There is evidence from a range of representative general population samples (4–9) as well as high-risk samples (10, 11) that low-grade psychotic experiences may precede affective and nonaffective clinical psychotic states by many years. Therefore, a dimensional focus, following the expression of different symptom domains and their relation to the onset of dysfunction and help-seeking over time, may be productive.
The aforementioned population-based work has focused on experiences resembling positive psychotic symptoms; it is not known if and how attenuated negative/disorganized symptoms may play a role over time. There is evidence that negative symptoms are expressed, in attenuated form, in otherwise “healthys” individuals and that the combination of attenuated positive and negative symptoms predicts onset of need for care (12).
In this article, we present a longitudinal, 10-year study of young people. Hypotheses were derived from the suggestion that negative/disorganized symptoms (13) associated with schizophrenia (14) may represent alterations in brain development (15–17) associated with distributed genetic risk (18), influencing a hypothesized final common pathway of neurotransmitter dysregulation (19), resulting in the onset of positive psychotic symptoms, particularly when combined with environmental exposures (20). It was hypothesized that 1) subclinical expressions of positive and negative/disorganized symptoms would be distributed in the general population and occur more often together than predicted by chance; 2) positive symptoms would be associated with environmental risks and negative/disorganized features with indices of developmental impairment; 3) negative/disorganized symptoms would predict psychotic experiences over time but not the other way around; and, finally, 4) “comorbid” expression of positive and negative/disorganized experiences would be associated with impairment and risk for patient status.
Method
Sample and Study Design
The Early Developmental Stages of Psychopathology Study collected data on the prevalence, incidence, risk factors, comorbidity, and course of mental disorders in a representative random population sample of adolescents and young adults, age 14–24 years, from the general population in Munich, Germany. The baseline sample, following ethics committee approval, was randomly drawn in 1994 from the respective population registry offices of Munich and its 29 counties to mirror the distribution of individuals expected to be 14–24 years of age at the time of the baseline (T0) interview in 1995. The base population were all those born between June 1, 1970, and May 31, 1981, registered as residents in these localities and having German citizenship. These registers can be regarded as highly accurate because 1) each German is registered by his town, 2) the registers are regularly updated, 3) in the interest of scientific studies, any number of randomly drawn addresses with a given sex and age group can be obtained, and 4) strict enforcement of registration by law and the police applies. More details on the sampling, representativeness, instruments, procedures, and statistical methods of the sample have previously been presented (21).
The longitudinal prospective design consisted of a baseline and three follow-up surveys (time 1, time 2, and time 3), covering a time period of on average 1.6, 3.5, and 8.4 years (range=7.3–10.5), respectively, from baseline. As the primary goal of the study was to examine developmental aspects of psychopathology, the younger group (14 and 15 years of age) was sampled at twice the rate of persons 16–21 years of age, and the oldest group (22–24 years) was sampled at half this rate. For the same reason, subjects 14–17 years of age at baseline were examined at all four time points, whereas the full baseline sample was only assessed at three time points (baseline, time 2, and time 3).
The present study is based on the three assessments that were available for the full baseline sample: 3,021 individuals 14–24 years of age at baseline and their follow-up assessments at time 2 (N=2,548; response rate=84%) and time 3 (N=2,210; 73%). Written informed consent was obtained from all participants.
Participants were interviewed with the Munich Composite International Diagnostic Interview (22). Trained and experienced clinical psychologists conducted the interviews at baseline (lifetime version) and follow-up assessments (interval version: covering the assessment period from the previous interview).
Psychopathology Assessment
Information from the psychosis section and the clinical interview rating section, with its embedded Brief Psychiatric Rating Scale (23), were used to derive measures of psychopathological clusters. In order to calculate measures of frequency, discrete variables indicating presence or absence of clusters across interview waves were, per definition, necessary.
Negative/disorganized symptom cluster.
A combined cluster of negative and disorganized symptoms was designed, in accordance with psychopathological findings in the early course of psychosis, showing that the “Bleulerian blend” of negative, catatonic/motor, and disorganization symptoms load on a single factor, while positive symptoms load on another (24) in agreement with Kraepelin's description in which avolition and disorganization were joined as the two defining features of the syndrome (25). Although negative, disorganized, positive, and affective clusters consistently load on separate domains in factor analytical studies of chronic psychotic patients (26, 27), meta-analysis indicates that negative and disorganized dimensions are conflated in their association with neurocognitive alterations, contrary to the positive and affective dimensions (28). A sensitivity analysis was planned, reanalyzing essential results for the negative cluster and the single disorganized item separately.
At baseline and time 3, two items concerning classic negative and disorganized symptoms from the interview ratings section were used: indifference (X11) and thought incoherence or illogicality (X12). These two items were rated on a 7-point scale (1=absent, 2=very little, 3=a little, 4=moderately, 5=moderately strong, 6=a lot, 7=very much). Each item was dichotomized: absent (coded as 0) versus present (coded as 1, indicating any score above “1”). For the purpose of the analyses, presence of the negative/disorganized cluster at baseline and time 3 was defined as a rating of “present” on either of the two items.
At time 2, not only the two previous items but also five items from the interview observation section were available: flat affect (P3), slow speech (P5), slow movement (P6), poverty of speech (P7), and avolition (P8). These five items were dichotomously rated: absent (coded as 0) versus present (coded as 1). Presence of the negative/disorganized cluster at time 2 was defined as a rating of “present” on any of the seven items.
Positive symptom cluster.
Interview ratings from the psychosis section (items G1, G2a, G3–G5, G7–G13, G13b, G14, G17, G18, G20, G20C, G21, and G22a) on delusions (15 items) and hallucinations (5 items) were collected at time 2 (lifetime) and time 3 (interval version). These 20 items concern classic psychotic symptoms, including Schneiderian first-rank symptoms such as audible thoughts, thought insertion, thought withdrawal, and made acts and impulses. Participants were first invited to read a list of all the psychotic experiences and asked whether they had ever experienced such symptoms. Each psychosis item was rated absent or present. Presence of the positive cluster was defined as a rating of “present” on any of the 20 psychosis items (6).
Depressive symptoms.
At time 2 and time 3, the 28 symptom items of the depression section were used. Items were rated either yes or no as being present for at least 2 weeks. A sum score of “depressive symptoms” was formed, with a minimum of 0 and a maximum score of 28 endorsements (29).
Cluster persistence.
The time 2 (lifetime) and time 3 (interval) ratings for the positive cluster were used to create a three-level summary “positive cluster persistence” variable: no occurrence (coded as 0), present at one assessment (coded as 1), and present at two assessments (coded as 2). In order to have a comparable longitudinal measure of “negative/disorganized cluster persistence,” the baseline (lifetime) and time 2 (interval) measures were combined into a cumulative lifetime variable. Together with the information at time 3 (interval), this resulted in a similar three-level summary score of persistence: no occurrence (coded as 0), present at one assessment (coded as 1), and present at two assessments (coded as 2)
Other Clinical Assessments
Age, gender, social status, marital status, level of education, urbanicity, cannabis exposure, and self-reported trauma were considered.
As part of the treatment module, participants were shown a list of different types of medication and were asked to endorse those they had been given for any psychopathological or psychosomatic problem. The acknowledgment of any antipsychotic medication (item Q1EA4) reported at time 2 and time 3 was used to derive the antipsychotic treatment variable (coded as 1 for yes).
Clinical relevance of positive psychotic symptoms was assessed with interview ratings of the psychosis section (6). Three help-seeking items assessed whether participants had ever sought help because of psychotic symptoms: seeking doctors' help because of delusions (G16) or hallucinations (G23) or seeking help from other mental health professionals, ranging from a general practitioner or school psychologist to psychiatric sheltered housing (Q1DG). A dichotomous “help-seeking” variable was created, indicating a positive answer to any of the three questions (coded as 1) versus negative answers on all.
The diagnostic interview also assesses the effect of psychotic experiences on feeling upset or unable to work, go places, or enjoy oneself at the time of having these experiences (G28); being less able to work (G29) or less able to make friends or enjoy social relationships (G29a) since these experiences began; and how much their life and everyday activities were impaired when these experiences were at their worst (G36). A dichotomous “dysfunction” variable was created representing a positive answer on any of the four questions (coded as 1) versus negative answers on all.
Finally, a dichotomous variable “psychotic impairment” was created, coded as 1 for subjects with psychotic experiences who had ratings of “1” on either help-seeking or dysfunction or both.
The interviewer's opinion regarding clinical evidence of psychological ill health (item X16) was rated according to four levels: 0=essentially not noticeable, 1=not very noticeable, 2=clearly ill, and 3=very ill. The dichotomous variable “caseness” indicated individuals with a noticeable level of psychiatric caseness (i.e., any score above “1”).
Statistical Analyses
All analyses were conducted using the software package STATA 10.1 (StataCorp, 2008).
First, in order to investigate the reliability and the rate of the variables representing the positive and the negative/disorganized psychopathological domains, the Cronbach alpha coefficient was used as a measure of internal consistency for each symptom cluster: the negative/disorganized cluster (24), the negative cluster (excluding the disorganized item), and the positive cluster (6) at time 2. Alpha values of 0.7 are regarded as satisfactory. Lifetime cumulative incidence and interval prevalence estimates of the positive cluster were calculated. Prevalence estimates of the negative/disorganized cluster were calculated, with and without additional sensitivity analyses excluding 1) participants exposed to antipsychotic treatment or 2) individuals with a DSM-IV diagnosis of depression (secondary negative symptoms). The distribution of symptom clusters was examined at time 2 and time 3, distinguishing individuals with only negative/disorganized symptoms, only positive symptoms, or a combination of the two. Because additional exclusion of participants exposed to antipsychotic treatment or individuals with a DSM-IV diagnosis of depression revealed no change in the rates, no sensitivity analysis was conducted.
Second, associations between psychopathology and risk factors were examined. It was hypothesized that the negative/disorganized cluster would be associated with younger age (30), male sex (31), and single marital status and low educational level (32), whereas female sex (31), cannabis exposure (33), urbanicity (34), self-reported trauma (34), and low level of education (32) would show significant associations with the positive cluster. Logistic regression analyses were conducted with the data in the “long” format, i.e., each individual contributing two timepoint observations (time 2 and time 3). Clustering within subjects was controlled for by including a variable reflecting measurement occasion in the model. Associations were expressed as odds ratios and their 95% confidence intervals (95% CI).
Third, in order to examine the natural association between the positive and negative/disorganized clusters (24), cross-sectional associations between the two were examined, controlling for depressive symptoms. Again, logistic regression analyses were carried out with the data in the “long” format (time 2 and time 3), adjusting for the indicator variable representing measurement occasion. Given that cross-sectional associations between symptoms may arise as a function of shared interview variance or represent chronicity effects, associations between clusters were also carried out in longitudinal predictive models across measurement occasions. Thus, logistic regression was used to calculate 1) the association between the negative/disorganized cluster at time 2 and the positive cluster at time 3, controlling for depressive symptoms and excluding individuals in whom the positive cluster was present at time 2; and 2) the association between the positive cluster at time 2 and the negative/disorganized cluster at time 3, controlling for depressive symptoms and excluding individuals in whom the lifetime negative/disorganized cluster was present at time 2.
Fourth, the null hypothesis of independence of the natural courses of the two psychopathological domains (35) was examined, analyzing the influence of the degree of persistence (from the interview assessing lifetime presence to time 3) of one psychosis cluster (i.e., positive cluster or negative/disorganized cluster) on time 3 onset of the other cluster and vice versa. To this end, logistic regression analyses were performed to examine the associations between 1) the degree of negative/disorganized persistence on the one hand and first onset (incidence) of positive cluster at time 3 on the other in individuals without lifetime positive cluster present at time 2, and 2) the degree of negative/disorganized persistence on the one hand and positive cluster at time 3 on the other in individuals with lifetime positive cluster present at time 2 and vice versa, controlling for depressive symptoms at time 3.
Fifth, the clinical relevance of the combination of positive and negative/disorganized experiences was calculated by examining whether the association between the positive cluster and its associated psychotic impairment rating was moderated by the negative/disorganized cluster. Consistent with previous work, the interaction between the positive cluster and the negative/disorganized cluster was calculated on the additive scale (33, 34, 36) using the STATA BINREG command, controlling for depression, with the data in the “long” format. In order to validate psychosis impairment, logistic regression analyses were conducted against the caseness and antipsychotic treatment variables.
Results
The average age of the participants at baseline was 18.3 years (range=14–24). Demographic characteristics are depicted in Table 1.
| Characteristic | N | % |
|---|---|---|
| Gender | ||
| Male | 1,533 | 50.74 |
| Female | 1,488 | 49.26 |
| Level of educationa | ||
| Low (mandatory basic school or learning a profession) | 475 | 15.72 |
| Medium (high school) | 915 | 30.29 |
| High (university or high school preparing for university) | 1,631 | 53.99 |
| Social statusb | ||
| Lower (lower class; lower middle class) | 207 | 6.85 |
| Middle (middle middle class) | 1,804 | 59.72 |
| Upper (higher middle class; upper class) | 950 | 31.45 |
| Other (none of the above or missing values) | 60 | 1.99 |
| Urbanicityc | ||
| Rural (surrounding areas of Munich) | 859 | 28.43 |
| Urban (city of Munich) | 2,162 | 71.57 |
TABLE 1. Demographic Characteristics of a General Population Sample of Adolescents and Young Adults (N=3,021) Followed Longitudinally for the Early Developmental Stages of Psychopathology Study
Natural Rates
Items in each psychopathological cluster were strongly correlated, yielding satisfactory internal consistency (seven negative/disorganized cluster items: Cronbach alpha=0.75; six negative cluster items [excluding the disorganized item]: Cronbach alpha=0.76; 20 positive cluster items: Cronbach alpha=0.67). The association between the disorganized item and negative cluster yielded an odds ratio of 13.9 (95% CI=6.5–29.6), against an odds ratio of 2.47 (95% CI=1.18–5.19) for the association between the disorganized item and positive cluster, supporting the a priori choice for a single negative/disorganized cluster.
The positive cluster cumulative lifetime incidence was 22.7% (N=574) at time 2 while the interval prevalence was 12.4% (N=274) at time 3. The baseline prevalence of the negative/disorganized cluster was 11.4% (N=345) while the follow-up prevalences were 12.8% (N=325) at time 2 and 12.2% (N=270) at time 3. No impact was seen on these rates following exclusion of individuals exposed to antipsychotic treatment (N=3 at time 2 and N=1 at time 3 for revised rates of 12.7% and 12.2%, respectively) or individuals with a DSM-IV diagnosis of depression (N=41 at baseline, N=43 at time 2, and N=32 at time 3 for revised rates of 11.3%, 12.3%, and 12.1%).
In terms of population distribution (Table 2), 15.7% of the 2,548 participants reported lifetime cumulative incidence of only negative/disorganized symptoms at time 2 and 17.0% reported only positive symptoms, whereas 5.5% reported both, more than twice what would be expected by chance alone (2.6%). Interval prevalence rates at time 3 were, respectively, 9.9%, 10.1%, and 2.3% of a total of 2,210 participants, again more than twice the expected rate (1%).
| Psychosis Variable and Timepoint | Total N | Negative/Disorganized Only | Positive Only | Co-Occurrence of Both Clusters | |||
|---|---|---|---|---|---|---|---|
| N | % | N | % | N | % | ||
| Symptom cluster rate | |||||||
| Time 2 (lifetime cumulative incidence) | 2,548 | 400 | 15.7 | 434 | 17.0 | 140 | 5.5 |
| Time 3 (interval prevalence) | 2,210 | 220 | 9.9 | 224 | 10.1 | 50 | 2.3 |
| Psychotic impairment rate | |||||||
| Time 2 (lifetime cumulative incidence) | 226 | 155 | 35.7b | 71 | 50.7b | ||
| Time 3 (interval prevalence) | 115 | 88 | 39.3c | 27 | 54.0c | ||
| Risk Difference (%)d | 95% CI | Risk Difference (%)d | 95% CI | Risk Difference (%)d | 95% CI | ||
| Excess risk for psychotic impairment | |||||||
| Time 2 | 0 | 37 | 32–41 | 53 | 43–64 | ||
| Time 3 | 0 | 39 | 33–46 | 54 | 39–67 | ||
| Both time 2 and time 3 | 0 | 37 | 34–41 | 53 | 45–62 | ||
TABLE 2. Psychopathological Clusters and Risk for Psychotic Impairmenta
Cluster Associations
The negative/disorganized cluster was significantly associated with younger age, male sex, single marital status, and low educational level, whereas the positive cluster was significantly associated with low educational level, cannabis exposure, urbanicity, and self-reported trauma. The combination of both clusters was significantly associated with all variables except gender (Table 3).
| Risk Factor | Negative/Disorganized Cluster | Positive Cluster | Co-occurrence of Both Symptom Clusters | |||
|---|---|---|---|---|---|---|
| Odds Ratio | 95% CI | Odds Ratio | 95% CI | Odds Ratio | 95% CI | |
| Age | 0.91** | 0.89–0.93 | 0.99 | 0.97–1.01 | 0.88** | 0.84–0.93 |
| Gender (male versus female) | 0.69** | 0.61–0.79 | 0.90 | 0.77–1.05 | 0.73 | 0.52–1.03 |
| Single marital status (never married, living apart, divorced, or widowed versus married) | 2.13** | 1.51–3.01 | 1.28 | 0.97–1.70 | 15.16** | 2.11–108.91 |
| Lower level of education (low versus medium and high) | 1.90** | 1.62–2.23 | 1.36** | 1.09–1.68 | 1.75* | 1.14–2.67 |
| Cannabis exposurea | 1.12 | 0.91–1.39 | 2.10** | 1.61–2.75 | 2.05** | 1.18–3.59 |
| Urbanicity (rural versus urban) | 1.12 | 0.97–1.29 | 1.19* | 1.01–1.41 | 1.51* | 1.01–2.27 |
| Traumab | 0.98 | 0.85–1.13 | 1.50** | 1.29–1.76 | 1.67** | 1.17–2.40 |
TABLE 3. Association of Sociodemographic and Environmental Risk Factors With Psychopathological Symptom Clusters at Time 2 (N=2,548) and Time 3 (N=2,210)
Cross-sectionally, the negative/disorganized cluster was significantly associated with the positive cluster, controlling for depressive symptoms (odds ratio=1.4, 95% CI=1.1–1.7). The prospective association between the negative/disorganized cluster at time 2 and the positive cluster at time 3, excluding individuals with presence of lifetime positive cluster at time 2 and controlling for depressive symptoms, was significant (odds ratio=2.3, 95% CI=1.5–3.7). In contrast, the prospective association between the positive cluster at time 2 and the negative/disorganized cluster at time 3, excluding individuals with the lifetime negative/disorganized cluster present at time 2 and controlling for depressive symptoms, was neither large nor significant (odds ratio=0.8, 95% CI=0.5–1.2).
Independence of Natural Course
After depressive symptoms were controlled, persistence of the negative/disorganized symptom cluster not only predicted positive cluster persistence but also time 3 incident psychotic experiences in a dose-response fashion. In contrast, persistence of the positive symptom cluster predicted neither negative/disorganized cluster persistence nor incident negative/disorganized symptoms at time 3 (Table 4).
| Exposure Variable: Symptom Cluster Persistencea | Outcome Variable: Incidence or Persistence of Symptoms | |||
|---|---|---|---|---|
| Odds Ratio | 95% CI | Odds Ratio | 95% CI | |
| Positive Symptom Cluster | Incident Negative/Disorganized Cluster | Persistent Negative/Disorganized Cluster | ||
| 0 | ||||
| 1 | 0.8 | 0.6–1.3 | 1.3 | 0.7–2.3 |
| 2 | 1.3 | 0.7–2.6 | 1.4 | 0.6–3.3 |
| Negative/Disorganized Symptom Cluster | Incident Positive Cluster | Persistent Positive Cluster | ||
| 0 | ||||
| 1 | 1.6* | 1.1–2.4 | 1.7* | 1.1–2.8 |
| 2 | 4.5** | 2.3–8.7 | 3.4** | 1.3–9.0 |
TABLE 4. Longitudinal Association of Symptom Cluster Courses
Clinical Relevance
Psychotic impairment was strongly associated with the variables of caseness (odds ratio=10.3, 95% CI=6.9–15.2) and antipsychotic treatment (odds ratio=15.3, 95% CI=6.1–38.4).
Table 2 depicts the proportions of individuals with psychotic impairment among those in whom the positive cluster was present at time 2 and time 3. Differences in risk for psychotic impairment for each psychopathological cluster were statistically significant in the combined time 2 and time 3 data analysis, as evidenced by the estimated additive interaction between the two clusters in the model of impairment (risk difference=16.2%, 95% CI=7.3%–25.2%). Analyzing time 2 and time 3 separately revealed replication at both time points (Table 2).
Separating Negative and Disorganized Symptoms
The rate of the separate negative cluster at the different time points was similar to that of the combined negative/disorganized cluster (prevalence rates around 12%), whereas the rate of the single item of disorganization was around 1%–3%. The pattern of associations with sociodemographic and environmental risk factors found for the combined negative/disorganized cluster was similar when analyzed separately for the negative cluster and disorganization item. Similarly, the cross-sectional associations between the positive cluster on the one hand and the separate negative cluster and disorganization item on the other was similar as observed for the original negative/disorganized cluster.
With regard to prospective associations, the negative cluster predicted onset and persistence of the positive cluster similar to the combined negative/disorganized cluster, whereas results for the single item of disorganization were directionally similar but less clear-cut (all results available upon request).
Discussion
Even after exclusion of ill individuals using antipsychotic medication, positive and negative/disorganized experiences resembling the symptoms of psychotic disorder occurred in the general population (5-year interval rates of around 12%). Positive and negative/disorganized experiences clustered together more often than predicted by chance, independent of depression, yet displayed a differential pattern of associations with sociodemographic and environmental risk factors. The negative/disorganized cluster was associated with a profile indicative of developmental impairment and the positive cluster with environmental risks. Their pattern over time revealed that persistent expression of the negative/disorganized cluster predicted onset of psychotic experiences; in contrast, persistent expression of psychotic experiences did not predict onset of negative/disorganized features. Finally, “comorbid” expression of the two clusters increased the risk for impairment and, by implication, need for care.
Negative/Disorganized Features Predict Positive Features and Onset of Clinical Psychosis
The negative/disorganized symptoms (13) associated with schizophrenia (14) may represent alterations in brain development (15–17) associated with distributed genetic risk (18), influencing a hypothesized final common pathway of neurotransmitter dysregulation (19) resulting in the onset of positive psychotic symptoms, particularly when combined with environmental exposures (20). Several elements in the current epidemiological analysis support this model.
First, the pattern of associations for the negative/disorganized cluster with young age, male sex, single marital status, and lower educational level resembles the pattern seen in clinical samples (30, 31). A profile indicative of developmental impairment is concordant with meta-analytic work demonstrating a unique association of the negative/disorganized domain with neurocognitive impairment (28), poorer social functioning (37), poorer response to antipsychotic treatment (38), and a worse prognosis (39).
Second, the sociodemographic associations of the positive cluster are in agreement with previous reports linking positive psychotic experiences to environmental risk factors such as cannabis exposure, urbanicity, self-reported trauma, and lower educational level (32–34).
Third, in most of the individuals both symptom clusters did not overlap. Nonetheless, the positive and negative/disorganized symptom domains clustered together more often than could be expected by chance alone. Thus, psychotic disorders such as schizophrenia may not be the result of the chance cross-section of the separate liabilities of developmental impairment on the one hand and reality distortion on the other; they are linked together in a way that suggests fundamental overlap or interaction of underlying mechanisms (Figure 1).
FIGURE 1. Negative/Disorganized Features Predict Positive Features and Onset of Clinical Psychosis
a The longitudinal analysis suggested that the negative/disorganized cluster associated with a sociodemographic pattern indicative of developmental impairment reflects an underlying vulnerability that, when combined with environmental exposures such as cannabis use, trauma, and urbanicity, results in positive psychotic symptoms increasing the risk of impairment and clinical relevance.
Fourth, importantly, the longitudinal analysis suggested that in this community sample negative/disorganized features—the negative features more than the disorganized—preceded positive features rather than the other way round. This finding is in agreement with follow-back studies showing that negative features frequently represent an initial symptom of schizophrenia, occurring more than 2 years before the emergence of positive symptoms (40).
Methodological Issues
First, assessment of psychotic experiences was based on self-report information by the respondents in diagnostic interviews, complemented by clinical ratings of core features of positive, negative, and disorganized symptoms derived from the clinical rating section and selected items from the Brief Psychiatric Rating Scale (23). False positive psychotic experiences are likely to have occurred, increasing random error; nevertheless, there is a substantial literature on the (predictive and other forms of) validity of self-reported psychotic experiences (1, 4–9).
Second, despite the hazard of under/overestimation of true positive, clinically relevant psychotic experiences (6), the validation of the outcome variable against caseness and antipsychotic treatment supports the validity of psychotic impairment as a proxy measure of clinical relevance.
Third, the definition of clinically relevant psychosis was contingent on the presence of positive psychotic experiences. This, however, is in line with the diagnostic criteria for schizophrenia, which require positive psychotic experiences for making the diagnosis.
Fourth, the measures of negative/disorganized symptoms were varying between measurement occasions. This is likely to have generated more random error resulting in a more conservative estimate of associations (i.e., the overlap between positive cluster and negative/disorganized cluster may be stronger than reported and the predictive value for positive psychotic experiences may be higher).
Finally, the threshold of the negative/disorganized cluster may be considered too low. However, a sensitivity analysis using a higher threshold (any score above “3”) yielded similar results (data not shown, available upon request).
1. : A systematic review and meta-analysis of the psychosis continuum: evidence for a psy-chosis proneness-persistence-impairment model of psychotic disorder. Psychol Med 2009; 39:179–195Crossref, Medline, Google Scholar
2. : Etiological and clinical features of childhood psychotic symptoms: results from a birth cohort. Arch Gen Psychiatry 2010; 67:328–338Crossref, Medline, Google Scholar
3. : Evidence that self-reported psychotic experiences represent the transitory developmental expres-sion of genetic liability to psychosis in the general population. Am J Med Genet B Neuropsychiatr Genet 2009; 150B:1078–1084Crossref, Medline, Google Scholar
4. : Emotional and behavioural antecedents of young adults who screen positive for non-affective psychosis: a 21-year birth cohort study. Psychol Med 2009; 39:625–634Crossref, Medline, Google Scholar
5. : Children's self-reported psychotic symptoms and adult schizophreniform disorder: a 15-year longitudinal study. Arch Gen Psychiatry 2000; 57:1053–1058Crossref, Medline, Google Scholar
6. : Evidence that onset of clinical psychosis is an outcome of progressively more persistent subclini-cal psychotic experiences: an 8-year cohort study. Schizophr Bull [Epub ahead of print: May 21, 2009]Google Scholar
7. : The incidence and outcome of subclinical psychotic experiences in the general population. Br J Clin Psy-chol 2005; 44(Pt 2):181–191Crossref, Medline, Google Scholar
8. : Self-reported psychotic symptoms in the community, and risk of later hospitalization for non-affective psychotic disorders, in Abstracts for the 12th International Congress on Schizophrenia Research (ICOSR). Schizophr Bull 2009; 35(March suppl):74Google Scholar
9. : Putatively psychosis-prone subjects 10 years later. J Abnorm Psychol 1994; 103:171–183Crossref, Medline, Google Scholar
10. : Prediction of psy-chosis in youth at high clinical risk: a multisite longitudinal study in North America. Arch Gen Psychiatry 2008; 65:28–37Crossref, Medline, Google Scholar
11. : Back to the future: predicting and reshaping the course of psychotic disorder. Arch Gen Psychiatry 2008; 65:25–27Crossref, Medline, Google Scholar
12. : Is our concept of schizophrenia influenced by Berkson's bias? Soc Psychiatry Psychiatr Epidemiol 2004; 39:600–605Crossref, Medline, Google Scholar
13. : The reliability of distinguishing primary versus secondary negative symptoms. Compr Psychiatry 1995; 36:421–427Crossref, Medline, Google Scholar
14. : A separate disease within the syndrome of schizophrenia. Arch Gen Psychiatry 2001; 58:165–171Crossref, Medline, Google Scholar
15. : Functional imaging of memory retrieval in deficit vs nondeficit schizophre-nia. Arch Gen Psychiatry 1999; 56:1117–1123Crossref, Medline, Google Scholar
16. : Structural abnormalities in deficit and nondeficit schizophrenia. Am J Psychiatry 1993; 150:59–65Link, Google Scholar
17. : White matter alterations in deficit schizophrenia. Neuropsychopharmacology 2009; 34:1514–1522Crossref, Medline, Google Scholar
18. : Widespread reductions of cortical thickness in schizophrenia and spectrum disorders and evidence of heritability. Arch Gen Psychiatry 2009; 66:467–477Crossref, Medline, Google Scholar
19. : From real-world events to psychosis: the emerging neuropharmacology of delusions. Schizophr Bull 2009; 35:668–674Crossref, Medline, Google Scholar
20. : Schizophrenia. Lancet 2009; 374(9690):635–645Crossref, Medline, Google Scholar
21. : The Early Developmental Stages of Psychopathology study (EDSP): a methodological update. Eur Addict Res 2000; 6:170–182Crossref, Medline, Google Scholar
22. : DIS-X-Interviews: Manual für Screening-Verfahren und Interview; Interviewheft Längsschnittuntersuchung (DIA-X-Lifetime); Er-gänzungsheft (DIA-X-Lifetime); Interviewheft Querschnittsuntersuchung (DIA-X-12. Monats-Version); Ergänzungsheft (DIA-X-12. Monats-Version); PC-Programm zur Durchführung der Interviews (Längsund Querschnittsuntersuchung); Auswertungsprogramm. Frankfurt, Germany, Swets & Zeitlinger, 1997Google Scholar
23. : The Brief Psychiatric Rating Scale. Psychol Rep 1962; 10:799–812Crossref, Google Scholar
24. : The dimensional structure of first episode psychosis: an exploratory factor analysis. Psychol Med 1998; 28:935–947Crossref, Medline, Google Scholar
25. : Demential Praecox and Paraphrenia. New York, Robert E Krieger, 1971Google Scholar
26. : The symptoms of chronic schizophrenia: a re-examination of the positive- negative dichotomy [see comments]. Br J Psychiatry 1987; 151:145–151Crossref, Medline, Google Scholar
27. : Symptomatic and neuropsychological components of defect states. Schizophr Bull 1985; 11:409–419Crossref, Medline, Google Scholar
28. : Are psychotic psychopathology and neurocognition orthogonal? a systematic re-view of their associations. Psychol Bull 2009; 135:157–171Crossref, Medline, Google Scholar
29. : Dimensions of depression, mania and psychosis in the general population. Psychol Med 2004; 34:1177–1186Crossref, Medline, Google Scholar
30. : The life course of schizophrenia: age and symptom dimensions. Schizophrenia Res 1997; 23:15–23Crossref, Medline, Google Scholar
31. : Sex differences in schizophrenia: a review of the literature. Acta Psychiatr Scand Suppl 2000; 401:3–38Crossref, Medline, Google Scholar
32. : Strauss (1969) revisited: a psychosis continuum in the general population? Schizophr Res 2000; 45(1–2):11–20Crossref, Medline, Google Scholar
33. : Prospective cohort study of cannabis use, predisposition for psychosis, and psychotic symptoms in young people. BMJ 2005; 330:11Crossref, Medline, Google Scholar
34. : Evidence that the outcome of developmental expression of psychosis is worse for adolescents growing up in an urban environment. Psychol Med 2006; 36:407–415Crossref, Medline, Google Scholar
35. : Structure and course of positive and negative symptoms in schizophrenia. Arch Gen Psychiatry 1995; 52:127–134Crossref, Medline, Google Scholar
36. : Biologic synergism and parallelism. Am J Epidemiol 1997; 145:661–668Crossref, Medline, Google Scholar
37. : The quantity and quality of the social networks of young people with early psychosis compared with closely matched controls. Schizophr Res 2000; 46:25–30Crossref, Medline, Google Scholar
38. : Treatment of schizophrenia negative symptoms: future prospects. Schizophr Bull 2006; 32:234–237Crossref, Medline, Google Scholar
39. : The positive-negative distinction in schizophrenia: review of natural history validators. Arch Gen Psychiatry 1992; 49:63–72Crossref, Medline, Google Scholar
40. : Early detection of schizophrenia: current evidence and future perspectives. World Psychiatry 2006; 5:130–138Medline, Google Scholar
