Abstract
Background:
Previously, we demonstrated that cognitive behavior therapy for ultra-high risk (called CBTuhr) halved the incidence of psychosis over an 18-month period. Follow-up data from the same study are used to evaluate the longer-term effects at 4 years post-baseline.
Method:
The Dutch Early Detection and Intervention Evaluation study was a randomized controlled trial of 196 UHR patients comparing CBTuhr with treatment-as-usual (TAU) for comorbid disorders with TAU only. Of the original 196 patients, 113 consented to a 4-year follow-up (57.7%; CBTuhr = 56 vs TAU = 57). Over the study period, psychosis incidence, remission from UHR status, and the effects of transition to psychosis were evaluated.
Results:
The number of participants in the CBTuhr group making the transition to psychosis increased from 10 at 18-month follow-up to 12 at 4-year follow-up whereas it did not change in the TAU group (n = 22); this still represents a clinically important (incidence rate ratio [IRR] = 12/22 = 0.55) and significant effect (F(1,5) = 8.09, P = .03), favoring CBTuhr. The odds ratio of CBTuhr compared to TAU was 0.44 (95% CI: 0.24–0.82) and the number needed to treat was 8. Moreover, significantly more patients remitted from their UHR status in the CBTuhr group (76.3%) compared with the TAU group (58.7%) [t(120) = 2.08, P = .04]. Importantly, transition to psychosis was associated with more severe psychopathology and social functioning at 4-year follow-up.
Conclusions:
CBTuhr to prevent a first episode of psychosis in persons at UHR of developing psychosis is still effective at 4-year follow-up. Our data also show that individuals meeting the formal criteria of a psychotic disorder have worse functional and social outcomes compared with non-transitioned cases. Trial registration: The trial is registered at Current Controlled Trials as trial number ISRCTN21353122 (http://controlled-trials.com/ISRCTN21353122/gaag).
Key words: schizophrenia, at-risk mental state, prevention, cognitive behavior therapy, ultra-high risk
Introduction
Since the establishment of ultra-high risk (UHR) criteria for psychosis,1 the aim is to prevent the onset of a first-episode of psychosis in UHR individuals. A recent meta-analysis of the effects of cognitive behavioral therapy (CBT) for young people at UHR showed encouraging results: CBT was associated with an overall risk reduction of transition to psychosis of 54% at 12-month follow-up, and 37% at 24–48 months.2 In addition to the clinical benefits of prevention of psychosis, healthcare costs may also be reduced by prevention of psychosis.3
In line with recent research,2,4–7 we demonstrated the efficacy of CBT for preventing psychosis in the Dutch Early Detection Intervention and Evaluation (EDIE-NL).8 Our previous 18-month follow-up results demonstrated a favorable effect on reducing the incidence of psychosis by about 50% compared to treatment-as-usual (TAU).8 This suggests that CBTuhr can delay or even prevent the onset of psychosis, and that CBT should be regarded as an evidence-based treatment in UHR symptoms.9
With the primary goal of preventing conversion to psychosis, the focus of the CBT intervention is mainly directed at the reduction of distress caused by positive symptoms and did not address other co-morbidities and social functioning. However, both conditions received TAU, which is guideline congruent care for comorbid psychopathology, expecting that in both conditions depression and anxiety will be reduced, but not per se as a consequence of offering CBT for preventing psychosis.
The threshold at which psychotic symptoms progress from sub-threshold to “psychotic disorder” is arbitrary; some patients with transitions to psychosis may recover quickly, while others have a bad outcome without transition to psychosis.10,11 Therefore, this study examines whether a first psychotic episode is associated with a worse clinical and social outcome in comparison to those not reaching threshold criteria.
This study examines the 4-year follow-up data from the Dutch EDIE-NL trial to evaluate: (1) whether CBTuhr has sustained preventive effects for first-episode of psychosis, (2) whether CBTuhr promotes remission from UHR status, and (3) the clinical and functional status of those who did convert to psychosis.
Methods
Setting and Sample
The Dutch EDIE-NL study was a randomized controlled trial with 196 patients at UHR of psychosis in secondary mental health care at 6 sites in the Netherlands.
Details of the study’s methodology are reported elsewhere.12 In brief, the inclusion criteria were: (1) age between 14–35 years; (2) a family history of psychosis or Comprehensive Assessment of At-Risk Mental States (CAARMS)1 scores in the range of the at-risk mental state (ARMS); and (3) an impairment in social functioning (a score on the “Social and Occupational Functioning Assessment Scale” [SOFAS]13 of 54 or less, and/or a reduction of 30% on the SOFAS for the duration of at least 1 month in the past year).
Exclusion criteria were: (1) the presence of a current or past psychotic disorder; (2) insufficient command of the Dutch language; (3) severe learning impairment; and (4) current or previous use of antipsychotic medication equivalent to a total cumulative haloperidol equivalent dose of ≥15mg.
The participants were randomized over 2 conditions. The random allocation list was separately generated for each site in random-permuted blocks of 10 by a web-based automated randomization system.8 Those who performed the research assessments were kept blind to randomization. In case the blinding was broken, the assessor was replaced when possible. This occurred 5 times.
Patients in both conditions received TAU, ie, routine care provided for the nonpsychotic disorders (mainly depressive and anxiety disorders) for which there were seeking treatment. This treatment was given according to the evidence-based clinical Dutch14 and the NICE guidelines.15 The experimental group received TAU plus CBTuhr,16 with a maximum of 26 sessions in the first 6 months after inclusion, with the aim to prevent first-episode psychosis.
Consenting participants were assessed at 4 years post-baseline. Research assistants contacted the participants between June 2012 and January 2014, requesting their participation in a 1-time interview regarding the longer-term course and outcome of (sub-threshold) psychotic symptoms. Participants were first contacted by telephone and asked if they would consent to a face-to-face interview. The interview consisted of psychopathological assessments and took place at local outpatient services for mental health care. If participants did not consent to a face-to-face interview, they were asked if they would consent to a brief telephone assessment, enabling the collection of a minimal set of clinical and functional outcome data (including all instruments mentioned under “Assessments” except the CAARMS subscales following after the positive symptoms).
Figure 1 presents the flowchart of the participants through the trial. Of the 201 patients included in the original trial, 5 participants (2.5%; 2 dissimulated a psychotic disorder and 3 had a history of psychosis) were included incorrectly at baseline (CBTuhr: 3 and TAU: 2) and were excluded from all further analyses in the current study.8 Assessors who were blind to randomization made the decisions. At the 18-month follow-up measurement, 140 participants (71.4%) participated and at 4 years 113 patients (57.7%) had a complete follow-up assessment. Of the 83 dropouts, 9 (10.8%) had already transitioned to psychosis during the first 18 months of the study and for 16 participants (19.3%) it could be verified that they were not diagnosed with a DSM psychotic disorder 4 years post-baseline, by reviewing the electronic health record (EHR), a digital system containing the medical history of participants in mental health care.
Fig. 1.
Flowchart of the study participants in the long-term follow-up trial.
Assessments
At baseline, demographic variables, clinical characteristics, and use of cannabis or any other substance abuse were recorded. The primary outcome of the current study was transition to psychosis. Transition was defined by the CAARMS using the recommended criteria of a global rating scale score of 6 on either unusual thought content, non-bizarre ideas, or disorganized speech, or 5–6 on perceptual abnormalities, with an associated frequency score of 4–6, for ≥1 week. Additionally, psychosis diagnosis was verified using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN).17 Measurements were performed at baseline and at 2, 4, 6, 9, 12, 15, and 18 months, or at the moment a therapist or research assistant informed the researchers that a transition had (probably) taken place. Details on the measurements and other baseline variables are reported elsewhere.12 The full 4-year follow-up assessment included:
(Sub-threshold) psychotic symptoms were assessed with the CAARMS.1 The CAARMS is a semi-structured interview that consists of 7 subscales, including 4 positive symptom items (unusual thought content, non-bizarre ideas, perceptual abnormalities, and disorganized speech), 2 cognitive and 3 emotional disturbance items, 3 negative symptom items, 4 behavioral change items, 4 motor/physical changes items, and 8 general psychopathology items. The global rating score (ie, symptom severity) and frequency of the symptoms is scored on a 0–6 Likert scale and distress caused by the symptom on a 0–100 scale. Symptomatic criteria for ARMS are based exclusively on positive symptom items. The EDIE-NL investigators received 2 days of training from Professor A. Yung, who developed the CAARMS criteria. All CAARMS interviews were discussed in a team of experienced CAARMS assessors.
The Social and Occupational Functioning Assessment Scale (SOFAS)13 was used to determine the level of social and occupational functioning; this scale ranges from 0 to 100.
Depression was assessed with the Beck Depression Inventory (BDI-2)18 and with the Calgary Depression Scale (CDS).19
The Social Interaction Anxiety Scale (SIAS) was used to measure self-reported social anxiety.20
The Manchester Short Assessment of Quality of Life21 measured perceived quality of life.
The Beliefs about Illness Questionnaire (PBIQ-R) was used to measure negative beliefs about the self as a patient.22,23
Healthcare usage was assessed with a short version of the Trimbos Institute and Institute of Medical Technology Assessment Questionnaire for Costs associated with Psychiatric Illness (TIC-P).24
The Schedules for Clinical Assessment in Neurop sychiatry (SCAN)17 to classify (according to a DSM-IV diagnosis) the patients who transitioned.
Data Analysis
The primary outcome was the difference in transition rates to psychosis at 4-year follow-up between the CBTuhr and TAU group. This was tested in a Poisson regression model to obtain incidence rate ratios (IRRs) and logistic regression analyses to obtain the ORs with STATA 13.1, using robust standard errors based on the first-order Taylor series linearization method for clustering,25 because the trial was conducted as a multisite trial. Missing clinical outcomes on the CAARMS at 4-year follow-up (n = 58) were imputed using the Expectation Maximization (EM) procedure as implemented in SPSS 22.0.0. The analysis were also repeated with Last Observation Carried Forward (LOCF) and Multiple Imputation (MI) as implemented in STATA 13.1, to ascertain that the findings did not solely hinge on the chosen imputation technique. For the EM and MI procedure we used baseline age, gender, ethnicity (Dutch or non-Dutch origin), baseline CAARMS distress positive symptoms and BDI scores as predictors. All predictor variables were identified as significant predictors of the imputed variable by multivariate regression analyses. Furthermore, we repeated the analysis on the sample without imputations and according to the intention-to-treat (ITT) principle, including 5 violations against the exclusion criteria.
The secondary aim was to investigate the difference in social functioning and psychopathology between the converters and non-converters. This analysis was performed in 113 participants of whom data were collected at 4-year follow-up. We present data on the standardized mean differences (SMDs) for baseline to 4-year follow-up for the transition group vs the non-transition group and the differential standardized mean differences between the groups, Δ(SMD).
To measure the effect sizes, standardized mean differences (Cohen’s) d for continuous outcomes were computed. Effect sizes of 0.8 can be assumed to be large, 0.5 moderate and 0.2 small.25
Results
Baseline Sample Characteristics
The baseline sample consisted of 95 participants in the CBTuhr group and 101 participants in the TAU group. There were no significant differences in demographic or clinical characteristics between the 2 groups (table 1).
Table 1.
Baseline Demographic and Clinical Characteristics of Participants in Both Conditions (CBTuhr vs TAU)
| CBTuhr (n = 95) | TAU (n = 101) | |
|---|---|---|
| Female gender, n (%) | 47 (49.5) | 52 (51.5) |
| Age on entry, mean (SD), years | 22.7 (5.6) | 22.6 (5.4) |
| Education, including university, mean (SD), years | 13.5 (3.8) | 13.2 (4.3) |
| Ethnicity minority, n (%) | 47 (49.5) | 40 (39.6) |
| Current marital status and living situation: single/ divorced, n (%) | 74 (77.9) | 79 (78.2) |
| Employment status, n (%) | ||
| Paid work | 43 (45.3) | 37 (36.6) |
| Unpaid work | 4 (4.2) | 9 (8.9) |
| School | 29 (30.5) | 29 (28.7) |
| Unemployment | 11 (11.6) | 19 (18.8) |
| Other | 8 (8.4) | 4 (4.0) |
| Clinical characteristics | ||
| BDI-II depression, mean (SD) | 20.8 (11.8) | 22.4 (12.9) |
| CDS depression, mean (SD) | 5.8 (4.8) | 6.4 (4.7) |
| SIAS anxiety, mean (SD) | 30.0 (16.5) | 31.2 (17.3) |
| CAARMS positive symptoms, mean (SD) | 10.2 (3.0) | 10.3 (2.6) |
| CAARMS negative symptoms, mean (SD) | 6.9 (3.3) | 7.4 (3.6) |
| CAARMS distress, mean (SD) | 173.1 (74.6) | 172.4 (75.2) |
| SOFAS social functioning, mean (SD) | 46.4 (4.8) | 45.6 (5.1) |
Note: BDI-II, Beck Depression Inventory II; CAARMS, Comprehensive Assessment of At-Risk Mental States; CBTuhr, cognitive behavior therapy for ultra-high risk; CDS, Calgary Depression Scale; SIAS, Social Anxiety and Interaction Scale; SOFAS, Social Functioning and Assessment Scale; TAU, treatment as usual.
Characteristics of the 4-Year Follow-up Cohort
Of the 113 (57.7%; CBTuhr: 56 and TAU: 57) participants that completed a 4-year follow-up assessment, 108 had an in-depth face-to-face interview, while 5 were interviewed via a brief telephone interview. A total of 83 participants could not be contacted or withdrew their original consent (figure 1). The mean follow-up time was 1520 (SD: 170.87; range: 1185–2439) days after baseline.
The 113 participants with a 4-year follow-up assessment were compared to the non-interviewed participants (n = 83) on clinical characteristics at the last observation, ie, the 18-month follow-up. In comparison to the interviewed, the non-interviewed participants showed less dysfunctional beliefs about their illness (d = −0.39; t(135) = −2.28, P = .02), fewer emotional disturbances (d = −0.40; t(135) = −2.30, P = .02), improved quality of life (d = −0.45; t(136) = 2.53, P = .01), and had lower symptom levels of general psychopathology (d = 0.40; t(135) = −1.91, P = .03; table 2).
Table 2.
Psychopathology Scores at 18-Month Follow-up of Participants and Nonparticipants at 4-Year Follow-up (N = 196)
| Participants (N = 113) | Nonparticipants (N = 83) | Effect Size, Cohen’s d | Test Statistic, P-value | |
|---|---|---|---|---|
| SOFAS, mean (SD) | 57.54 (13.48), N = 105 | 58.86 (14.16), N = 63 | d = −0.09 | t(166) = 0.60, P = .55 |
| BDI-II depression, mean (SD) | 11.30 (10.14), N = 86 | 9.00 (10.32), N = 50 | d = −0.22 | t(134) = −1.27, P = .21 |
| CDS depression, mean (SD) | 4.13 (4.62), N = 105 | 4.07 (5.40), N = 56 | d = −0.01 | t(159) = −0.08, P = .94 |
| SIAS anxiety, mean (SD) | 22.16 (14.18), N = 86 | 19.64 (14.34), N = 50 | d = −0.17 | t(134) = −1.00, P = .32 |
| PBIQ-R dysfunctional beliefs, mean (SD) | 62.70 (16.66), N = 87 | 55.58 (19.18), N = 50 | d = −0.39 | t(135) = −2.28, P = .02 |
| CAARMS positive symptoms, mean (SD) | 6.22 (5.32), N = 102 | 5.82 (4.81), N = 66 | d = −0.08 | t(166) = −0.49, P = .62 |
| CAARMS negative symptoms, mean (SD) | 3.50 (3.41), N = 88 | 2.51 (2.70), N = 49 | d = −0.30 | t(135) = −1.75, P = .08 |
| CAARMS distress, mean (SD) | 100.18 (98.10), N = 102 | 81.97 (89.10), N = 66 | d = −0.19 | t(166) = −1.22, P = .23 |
| CAARMS cognitive symptoms, mean (SD) | 2.24 (1.66), N = 89 | 1.80 (1.61), N = 49 | d = −0.26 | t(136) = −1.51, P = .13 |
| CAARMS emotional disturbance symptoms, mean (SD) | 1.67 (2.26), N = 88 | 0.84 (1.56), N = 49 | d = −0.40 | t(135) = −2.30, P = .02 |
| CAARMS behavioural symptoms, mean (SD) | 3.74 (3.77), N = 88 | 3.12 (3.53), N = 49 | d = −0.16 | t(135) = −0.94, P = .35 |
| CAARMS motor/physical symptoms, mean (SD) | 1.67 (2.13), N = 88 | 1.39 (3.04), N = 49 | d = −0.11 | t(135) = −0.64, P = .53 |
| CAARMS general psychopathology, mean (SD) | 7.77 (6.27), N = 88 | 5.31 (6.20), N = 49 | d = −0.40 | t(135) = −2.22, P = .03 |
| CAARMS total, mean (SD) | 25.20 (18.69), N = 88 | 19.08 (16.57), N = 49 | d = −0.35 | t(135) = −1.91, P = .06 |
| MANSA quality of life, mean (SD) | 53.85 (13.85), N = 88 | 59.78 (12.01), N = 50 | d = 0.45 | t(136)=2.53, P = .01 |
Note: MANSA, Manchester Short Assessment of Quality of Life; PBIQ-R, Personal Beliefs About Illness Questionnaire Revised. Bold numbers indicate significant differences (P < .05).
Long-term Impact on Psychosis Incidence
As shown in table 3, the number of participants converting to psychosis in the CBTuhr group increased from 10 at 18 months to 12 at 4-year follow-up. In the TAU group, the number of participants who transitioned to psychosis at 4-year follow-up remained the same (ie, 22). The number of days to transition to psychosis was significantly higher in the CBTuhr group (1322.45 days, SE: 38.99, 95% CI = 1246.03–1398.88) compared with the TAU group (1188.91, SE: 51.48, 95% CI = 1088.02–1289.81) (d = 0.29; t(194) = 2.03, P = .04).
Table 3.
Number and Characteristics of Transitions to First-Episode Psychosis for CBTuhr vs TAU at 4-Year Follow-up
| Condition | Number of Transitions to Psychosis (LOCF) | Number of Transition to Psychosis (EM) | Number of Transition to Psychosis (MI) | Days to Transition, Mean (SD) | Admission to Psychiatric Hospitala | Total Number of Months in Psychiatric Hospital From 18 to 48 Monthsb | Remission of Subclinical Psychotic Symptomsa |
|---|---|---|---|---|---|---|---|
| CBTuhr | 12/95 | 12/95 | 15/95 | 1322.45 (38.99), 95% CI: 1246.03–1398.88 | 4 patients with 5 admissions | 11.5 | 45/59 |
| TAU | 22/101 | 25/101 | 27/101 | 1188.91 (51.48), 95% CI: 1088.02–1289.81 | 7 patients with 15 admissions | 33 | 37/63 |
Note: DSM-IV, Diagnostic and Statistical Manual of Mental Disorders; EM, Expectation Maximization; Fourth Edition; LOCF, Last Observation Carried Forward; MI, Multiple Imputation.
aBased on the 113 participants that were followed-up at 4 years and the 9 participants that had transitioned to a first episode of psychosis between 0–18 months but could not be followed-up at 4 years post-baseline.
bCBTuhr, n = 56; TAU, n = 57.
A Poisson regression analysis on the EM-imputed sample revealed a significant effect (IRR = 0.51; 95% CI = 0.28–0.94) between the 4-year transition rates for the CBTuhr (12.6%) and TAU conditions (24.8%), favoring CBTuhr (F(1,5) = 8.09, P = .03). The OR of the corresponding logistic regression analysis was 0.44 (95% CI = 0.24–0.82; t(196) = −3.39, P = .02) and the number needed to treat (NNT) was 8.25 (95% CI = 4.37–72.72). Using the LOCF and the MI imputation, the Poisson regressions also revealed significant effects (LOCF: IRR = 0.58; 95% CI = 0.35–0.95 and MI: IRR = 0.56; 95% CI = 0.37–0.85) between the 4-year transition rates for both conditions, favoring CBTuhr in both scenario’s (LOCF: F(1,5) = 7.98, P = .04 and MI: F(1,5) = 27.74, P < .01). The OR of the logistic regression analysis for the EM was 0.52 (95% CI = 0.31–0.87), which was significant (t(196) = 3.30, P = .02) and 2.05 (95% CI = 1.41–2.96) for the MI imputation (t(196) = 4.97, P < .01). A Poisson regression based on the sample without imputation showed a significant difference between both conditions (IRR = 0.56; 95% CI = 0.37–0.85, favoring CBTuhr (F(1,5) = 13.14, P = .02). The OR of the corresponding logistic regression analysis was 0.47 (95% CI = 0.29–0.74), which was significant (t(138) = −4.21, P = .01). An ITT Poisson regression based on EM imputation, did not show a significant difference between CBTuhr and TAU (IRR: 0.58 [95% CI = 0.27–1.26]; F(1,5) = 4.24, P = .10). The OR of the corresponding logistic regression analysis was 0.51 (95% CI = 0.22–1.18), which was not significant (t(201) = −2.06, P = .10).
Participants in the TAU group were more often admitted to a psychiatric hospital (TAU: 7 participants who had a total of 15 admissions vs CBTuhr: 4 participants who had 5 admissions). The total number of months spent in hospital was 11.5 (CBTuhr) vs 33.0 (TAU); this difference was not significant [t(109) = −1.006, P = .32].
According to our second hypothesis, UHR status was examined at 4-year follow-up. In the sample with a complete 4-year follow-up measurement, there were significantly more participants in remission of UHR status in the CBTuhr group (76.3%) compared with the TAU group (58.7%) [t(120) = 2.08, P = .04]. The OR was 0.44 (95% CI = 0.20–0.97).
At 4-year follow-up, there were no significant differences in psychopathology in the group of non-converters for both treatment conditions (CBTuhr with TAU; table 4).
Table 4.
Psychopathology Scores for the Non-converters in Both Conditions (CBTuhr vs TAU) at Baseline, 18 and 48 Months Follow-up
| Baseline | 18 Months F-U | 48 Months F-U | |||||
|---|---|---|---|---|---|---|---|
| Clinical Characteristics | CBTuhr (n = 95) | TAU (n = 101) | CBTuhr (n = 71) | TAU (n = 69) | CBTuhr (n = 47) | TAU (n = 41) | Effect Size and Test |
| BDI-II depression, mean (SD) | 20.8 (11.8) | 22.4 (12.9) | 9.6 (9.4) | 11.3 (11.1) | 10.5 (8.7) | 12.2 (11.8) | d = 0.17, t(75) = −0.72, P = .47 |
| CDS depression, mean (SD) | 5.8 (4.8) | 6.4 (4.7) | 2.6 (3.7) | 3.3 (4.4) | 3.0 (3.8) | 3.7 (4.1) | d = −0.19, t(81) = −0.86, P = .40 |
| SIAS anxiety, mean (SD) | 31.0 (16.5) | 32.2 (17.3) | 21.2 (13.8) | 21.3 (15.2) | 18.7 (13.8) | 20.6 (17.2) | d = −0.12, t(83) = −0.57, P = .57 |
| CAARMS positive symptoms, mean (SD) | 10.2 (3.0) | 10.3 (2.5) | 4.1 (4.2) | 4.9 (3.5) | 2.6 (2.9) | 3.6 (3.1) | d = −0.33, t(86) = −1.54, P = .13 |
| CAARMS negative symptoms, mean (SD) | 6.9 (3.3) | 7.4 (3.6) | 2.8 (2.8) | 3.5 (3.6) | 3.6 (3.5) | 3.2 (3.2) | d = 0.12, t(83) = 0.56, P = .58 |
| CAARMS distress, mean (SD) | 173.1 (74.6) | 171.0 (75.2) | 71.9 (88.9) | 73.9 (78.2) | 55.0 (65.3) | 78.5 (86.9) | d = −0.31, t(82) = −1.41, P = .16 |
| SOFAS social functioning, mean (SD) | 46.4 (4.8) | 45.6 (5.1) | 61.6 (12.8) | 59.6 (13.7) | 59.0 (11.5) | 60.2 (14.4) | d = −0.10, t(86) = −0.46, P = .65 |
Note: F-U, follow-up.
Differences in Psychopathology and Social Functioning for the Transition vs the Non-transition Group From Baseline to 4-Year Follow-up
There were no differences in marital status and employment status between the transition and non-transition group.
Table 5 presents data on the standardized mean differences for baseline to 4-year follow-up for the transition group vs the non-transition group. We look at the within-group standardized mean difference scores over time, SMD, that reflect the longitudinal change in each group, and the between-group difference to see what group made the largest change, denoted Δ(SMD). The Δ(SMD), was −0.24 for positive and also for negative CAARMS symptoms, indicating that the transition group changed for the worse in positive and also in negative symptoms by 0.24 standard units relative to the group that did not transition. The Δ(SMD) for cognitive symptoms was −0.41 and also for CAARMS distress, and the Δ(SMD) for CAARMS behavioral change was −0.14, and for SIAS anxiety −0.07. All indicating that the transition group changed for the worse. Furthermore, the differential SMD for the SOFAS was 1.43; since the SOFAS is a scale with higher scores indicating a better outcome, this means that the transitioned group changed for the worst by 1.43 standard units compared to the non-transitioned group.
Table 5.
Effects of Transition on Clinical and Social Measures (n = 113)
| Transitions (n = 25) | Non-transitions (n = 88) | Between Groups | |||||
|---|---|---|---|---|---|---|---|
| Clinical Measures | Baseline Mean (SD) | 4-year F-U Mean (SD) | SMD (Change) | Baseline Mean (SD) | 4-Year F-U Mean (SD) | SMD (Change) | Δ(SMD) |
| CAARMS positive symptoms (0–24) | 8.44 (3.50) | 5.90 (3.80) | −0.73 | 7.10 (3.80) | 3.40 (3.40) | −0.97 | −0.24 |
| CAARMS negative symptoms (0–18) | 8.44 (3.50) | 5.90 (3.80) | −0.73 | 7.10 (3.80) | 3.40 (3.40) | −0.97 | −0.24 |
| CAARMS cognitive symptoms (0–12) | 4.12 (2.00) | 3.40 (1.80) | −0.36 | 3.50 (2.20) | 1.80 (1.30) | −0.77 | −0.41 |
| CAARMS emotional disturbance (0–18) | 8.44 (3.50) | 3.10 (2.60) | −1.53 | 3.56 (2.50) | 1.25 (2.10) | −0.92 | 0.61 |
| CAARMS behavioral change (0–24) | 10.56 (3.50) | 7.30 (4.30) | −0.93 | 9.30 (4.40) | 4.60 (4.10) | −1.07 | −0.14 |
| CAARMS motor/ physical change (0–24) | 4.80 (3.30) | 2.10 (1.90) | −0.82 | 3.70 (3.40) | 1.40 (1.60) | −0.68 | 0.14 |
| CAARMS general psychopathology (0–48) | 17.80 (5.40) | 10.40 (5.60) | −1.37 | 14.30 (6.60) | 7.60 (5.70) | −1.02 | 0.35 |
| CAARMS distress (0–100) | 207.80 (71.50) | 123.00 (97.40) | −1.19 | 173.40 (67.30) | 65.90 (76.50) | −1.60 | −0.41 |
| SOFAS, social functioning (0–100) | 44.50 (4.60) | 51.00 (11.90) | 1.41 | 47.10 (4.40) | 59.60 (12.90) | 2.84 | 1.43 |
| BDI-II depression (0–63) | 27.40 (10.80) | 15.90 (10.20) | −1.06 | 22.12 (13.70) | 11.30 (10.20) | −0.79 | 0.27 |
| CDS depression (0–39) | 8.56 (3.90) | 5.90 (5.30) | −0.68 | 5.80 (4.90) | 3.30 (3.90) | −0.51 | 0.17 |
| SIAS anxiety (0–80) | 35.80 (12.60) | 27.70 (16.10) | −0.64 | 31.90 (17.40) | 19.60 (15.40) | −0.71 | −0.07 |
Note: SMD, standardized mean difference.
Discussion
Main Findings
The present study demonstrates that the preventive CBTuhr intervention was successful in halving the 4-year cumulative incidence rate of psychosis in UHR participants (P = .03) compared to TAU. The favorable long-term outcomes of CBTuhr were also reflected in the low NNT, indicating that a psychosis can be prevented in every 8 UHR participants who receive the CBTuhr intervention instead of TAU. This can be improved with more accurate at risk criteria. The cost-effectiveness of this preventive CBTuhr intervention up to 18 months has been reported elsewhere3 and longer-term follow-up is currently being evaluated.
More participants had remitted from their sub-threshold psychotic symptoms in the CBTuhr group relative to the TAU group (P = .04). Remission from sub-threshold psychotic symptoms was reported in 76.3% (45/59) of the participants who received CBTuhr and in 58.7% (37/63) of the participants receiving TAU. Remission from UHR status is clinically relevant, as 50% of the transitions from UHR to psychosis occur within the first 12 months after clinical presentation,8 and persisting or recurrent sub-threshold positive psychotic symptoms may flag up an elevated risk for psychosis since transitions can occur up to 10 years after baseline, albeit in a small proportion.26 Comparable to our 18-months results,8 depression and anxiety scores were reduced at the follow-up assessments but still present. Thus, the focus on transition must be broadened as the UHR group who does not transition to psychosis is still not functioning well and suffering from a broad range of psychopathology and limitations in social functioning.9 In future intervention studies, the focus must rely on a broader set of outcomes as the UHR group is not only psychosis-prone, but more general psychopathology-prone.
Moreover, patients who transitioned to a psychotic disorder had a poorer outcome with respect to psychopathology and social functioning scores compared to those who did not convert.
Limitations
The present study has several methodological limitations that need to be considered, most of which are already documented.8 First, the most important limitation was a considerable loss to the 4-year follow-up (42.4%); this might be because participants experienced an increase in their mental health problems, or transition to psychosis. However, since the participants lost to follow-up had significantly less psychopathology, fewer dysfunctional beliefs about their illness and experienced better quality of life at the 18-month follow-up, it is more likely that they were the patients who improved and, for that reason, were no longer willing to participate. Some patients informed us that they were now functioning well and preferred not to be reminded about that particular episode in their lives. Another reason for attrition could be the mobility of this relatively young population, eg, moving house without providing a new address and/or frequent change of mobile telephone numbers. Furthermore, apart from age [the non-interviewed sample was younger (mean = 21.58, SE = 5.46 y) than the interviewed sample (mean = 23.45, SE = 5.37 y) [d = −0.34; t(194) = −2.39, P = .02]], there were no differences between the groups in relation to dropout.
Second, loss to follow-up may have caused an underestimation of the transition rate at 4-year follow-up. To establish the transition status for participants who were unavailable for interview, we reviewed the EHRs; this was possible in 16 of 74 cases. However, because these EHRs only contained information from the health service providing the initial treatment (ie, CBTuhr or TAU), this means that our participants could have been in contact with other mental health services. Also, there is a possibility that some transitioned cases were not detected if, for instance, no mental health service was in contact with the patient after transition to psychosis. Therefore, missing endpoints were replaced by their most likely value, using 3 different imputation strategies (EM, LOCF and MI imputation) to ascertain that the findings did not solely hinge on the imputation technique selected. However because all analyses yielded the same results, this underscores the robustness of our findings.
Third, the small number of people that made the transition to psychosis would warrant a larger sample as to draw firmer conclusions. Therefore, we recommend that replication studies must be conducted and that their data be combined meta-analytically as to overcome power issues. That said, the transition rates we observed were comparable to those in other trials27 but decreased over time following the index year, which may reflect a true trend of younger people being at higher risk of transition and the risk rates getting smaller when people mature into adulthood.27
Fourth, within the first 18 months of the study, all participants who transitioned to psychosis were not followed up in the same way as those who did not. As a consequence, some outcome measures were not available for converters at 4-year follow-up and therefore we were not able to report on measures such as the Psychotic Symptom Rating Scales (PSYRATS)28 and the Positive and Negative Syndrome Scale (PANSS).29
Fifth, both the experimental and the control group received treatment for their nonpsychotic disorders. If reducing the emotional distress as a consequence of therapy would also prevent transition to psychosis, then we would expect to find no difference between the 2 conditions. However, because an effect was found, this indicates that treating the current disorder may be beneficial, and that our CBTuhr has a specific effect over and above effects that were possibly stemming from TAU on treating subclinical psychotic symptoms and preventing transition to psychosis.
Sixth, we presented full ITT analyses that showed the significant difference between the groups was not significant anymore when including the 5 violations against the inclusion criteria.
Seventh, adverse events were not reported in a systematic manner. However, suicidality and self-harm were administered as a part of the CAARMS interview. We are not aware of any adverse events like suicide in the participants that were not seen for the 4-year follow-up.
Eight, participants and routine care clinicians were not blind for randomization status. This could have biased the findings.
Finally, in some cases the 95% CIs included trivial and very large effects, probably due to limited statistical power. Therefore the results have to be interpreted with caution.
Implications of Preventing Transition of a First Episode of Psychosis
Our data show that transition to a psychotic disorder is associated with a range of poor clinical and social outcomes in many domains in comparison to those not reaching threshold criteria. These findings add to the ongoing debate in the UHR field about the purpose and clinical use of the arbitrary transition criteria.10,11,30 In the present study, in participants who experienced a first-episode of psychosis, the change of positive, negative, cognitive, as well as behavioral change symptoms and anxiety and social functioning was worse in comparison to the non-transitioned group. Therefore, our findings support the notion that the prevention of a first episode of psychosis is an important goal to pursue.
In our study, the highest risk of transition to psychosis was present within the first 18 months and, in particular, within the first 12 months; this temporal pattern of transition is comparable to the findings of a recent meta-analysis.31 Also, at 4-year follow-up, only 2 more patients had transitioned to psychosis since the 18-month assessment. This result is in line with a recent 10-year follow-up study, in which 86% of the transitions took place in the first 5 years,26 suggesting that there is a critical period (up to early adulthood) with the highest risk in the first 12 months after established UHR status,8 but not a lifelong threat. Our data suggest that offering a cost-effective CBTuhr intervention3 in the first half year after detection of UHR status, and further monitoring of persisting or reoccurring UHR symptoms in the first 5 years after detection, could have a durable effect on outcome. Nevertheless, delaying the onset of a first episode of psychosis even for a few years could make a substantial difference in outcomes that impact on the patients and their families, as well as on public health costs.
Conclusion
This trial shows that CBT for patients at UHR of psychosis was successful in reducing the risk of a first psychosis by 50% and that these favorable effects were sustained over 4 years. In addition, the CBT intervention achieved higher remission rates in UHR symptomatology. Delay of the onset of a first psychotic episode might prove to be more than postponement of transition to psychosis alone and might be regarded as real prevention, since very few participants made the transition from UHR to psychosis after the first 12 (and 18) months. However, social functioning is still affected in most participants (even in participants that remitted from UHR status) and many still suffer from positive, negative, cognitive symptoms, behavioral change symptoms and anxiety. Much is gained by preventing a first psychotic episode, since patients who made a transition showed more severe psychopathology and worse social functioning compared to those who did not transition. The prognosis of the UHR status is much more favorable than the prognosis after a transition to a first psychotic episode.
We conclude that the prognosis of UHR patients appears to improve substantially by intervening with CBTuhr during a critical period and that this effect is sustained over many years; this beneficial effect probably prevents first psychotic episodes in a considerable number of patients.
Funding
The Netherlands Organization for Health Research and Development (ZonMw) supported this study (grant number 120510001). ZonMw had no further role in the study design, collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.
Acknowledgments
The authors are grateful for the valuable contributions made by all participants, research assistants, therapists, and all others who participated in or contributed to the EDIE-NL study. We also thank Ms Marion Bruns for preparation and organization related to the study. All authors confirm there are no conflicting interests.
References
- 1. Yung AR, Yuen HP, McGorry PD, et al. Mapping the onset of psychosis: the Comprehensive Assessment of At-Risk Mental States. Aust N Z J Psychiatry. 2005;39:964–971. [DOI] [PubMed] [Google Scholar]
- 2. van der Gaag M, Smit F, Bechdolf A, et al. Preventing a first episode of psychosis: meta-analysis of randomized controlled prevention trials of 12 month and longer-term follow-ups. Schizophr Res. 2013;149:56–62. [DOI] [PubMed] [Google Scholar]
- 3. Ising HK, Smit F, Veling W, et al. Cost-effectiveness of preventing first-episode psychosis in ultra-high-risk subjects: multi-centre randomized controlled trial. Psychol Med. 2015;45:1435–1446. [DOI] [PubMed] [Google Scholar]
- 4. Fusar-Poli P, Borgwardt S, Bechdolf A, et al. The psychosis high-risk state: a comprehensive state-of-the-art review. JAMA Psychiatry. 2013;70:107–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Hutton P, Taylor PJ. Cognitive behavioural therapy for psychosis prevention: a systematic review and meta-analysis. Psychol Med. 2014;44:449–468. [DOI] [PubMed] [Google Scholar]
- 6. Stafford MR, Jackson H, Mayo-Wilson E, Morrison AP, Kendall T. Early interventions to prevent psychosis: systematic review and meta-analysis. BMJ. 2013;346:f185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Thompson E, Millman ZB, Okuzawa N, et al. Evidence-based early interventions for individuals at clinical high risk for psychosis: a review of treatment components. J Nerv Ment Dis. 2015;203:342–351. [DOI] [PubMed] [Google Scholar]
- 8. van der Gaag M, Nieman DH, Rietdijk J, et al. Cognitive behavioral therapy for subjects at ultrahigh risk for developing psychosis: a randomized controlled clinical trial. Schizophr Bull. 2012;38:1180–1188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Schmidt SJ, Schultze-Lutter F, Schimmelmann BG, et al. EPA guidance on the early intervention in clinical high risk states of psychoses. Eur Psychiatry. 2015;30:388–404. doi:10.1016/j.eurpsy.2015.01.013. [DOI] [PubMed] [Google Scholar]
- 10. Lin A, Nelson B, Yung AR. “At-risk” for psychosis research: where are we heading? Epidemiol Psychiatr Sci. 2012;21:329–334. doi:10.1017/S2045796012000388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Yung AR, Nelson B, Thompson A, Wood SJ. The psychosis threshold in Ultra High Risk (prodromal) research: is it valid? Schizophr Res. 2010;120:1–6. doi:10.1016/j.schres.2010.03.014. [DOI] [PubMed] [Google Scholar]
- 12. Rietdijk J, Dragt S, Klaassen R, et al. A single blind randomized controlled trial of cognitive behavioural therapy in a help-seeking population with an At Risk Mental State for psychosis: the Dutch Early Detection and Intervention Evaluation (EDIE-NL) trial. Trials. 2010;11:30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Goldman HH, Skodol AE, Lave TR. Revising axis V for DSM-IV: a review of measures of social functioning. Am J Psychiatry. 1992;149:1148–1156. [DOI] [PubMed] [Google Scholar]
- 14.Trimbos Institute (Netherlands Institute of Mental Health and Addiction). Landelijke Stuurgroep Multidisciplinaire Richtlijnontwikkeling in de GGZ. [Dutch Committee for the Development of Multidisciplinary Guidelines in Mental Health Utrecht, The Netherlands: Trimbos Institute; 2008. www.ggzrichtlijnen.nl. Accessed March 7, 2016. [Google Scholar]
- 15. National Institute for Health and Care Excellence. Psychosis and Schizophrenia in Children and Young People: Recognition and Management (CG155). London, UK: National Institute for Health and Care Excellence; 2013. [Google Scholar]
- 16.van der Gaag M, Nieman D, van den Berg D. CBT for Those at Risk of a First Episode Psychosis: Evidence-based Psychotherapy for People With an “At Risk Mental State.” London, UK: Routledge; 2013. [Google Scholar]
- 17. Rijnders CA, van den Berg JF, Hodiamont PP, et al. Psychometric properties of the schedules for clinical assessment in neuropsychiatry (SCAN-2.1). Soc Psychiatry Psychiatr Epidemiol. 2000;35:348–352. [DOI] [PubMed] [Google Scholar]
- 18. Beck A, Steer R, Brown G. Manual for Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation; 1996. [Google Scholar]
- 19. Addington D, Addington J, Maticka-Tyndale E, Joyce J. Reliability and validity of a depression rating scale for schizophrenics. Schizophr Res. 1992;6:201–208. [DOI] [PubMed] [Google Scholar]
- 20. Mattick RP, Clarke JC. Development and validation of measures of social phobia scrutiny fear and social interaction anxiety11Editor’s note: this article was written before the development of some contemporary measures of social phobia, such as the Social Phobia and Anxiety Inve. Behav Res Ther. 1998;36:455–470. [DOI] [PubMed] [Google Scholar]
- 21. Priebe S, Huxley P, Knight S, Evans S. Application and results of the Manchester Short Assessment of Quality of Life (MANSA). Int J Soc Psychiatry. 1999;45:7–12. [DOI] [PubMed] [Google Scholar]
- 22. Birchwood M, Jackson C, Brunet K, Holden J, Barton K. Personal beliefs about illness questionnaire-revised (PBIQ-R): reliability and validation in a first episode sample. Br J Clin Psychol. 2012;51:448–458. [DOI] [PubMed] [Google Scholar]
- 23. Taylor PJ, Pyle M, Schwannauer M, Hutton P, Morrison A. Confirming the structure of negative beliefs about psychosis and bipolar disorder: a confirmatory factor analysis study of the Personal Beliefs about Experience Questionnaire and Personal Beliefs about Illness Questionnaire. Br J Clin Psychol. 2015;54:361–377. doi:10.1111/bjc.12079. [DOI] [PubMed] [Google Scholar]
- 24. Hakkaart-van Rooijen L. Manual Trimbos/iMTA Questionnaire for Costs Associated With Psychiatric Illness (in Dutch). Rotterdam, the Netherlands: Institute for Medical Technology Assessment, Erasmus University; 2002. [Google Scholar]
- 25.Lavrakas PJ. Encyclopedia of Survey Research Methods. Thousand Oaks, CA: SAGE Publications, Inc; 2008. [Google Scholar]
- 26.Nelson B, Yuen HP, Wood SJ, et al. Long-term follow-up of a group at ultra high risk (“prodromal”) for psychosis: the PACE 400 study. JAMA Psychiatry. 2013:70:793–802. [DOI] [PubMed] [Google Scholar]
- 27. Fusar-Poli P, Bonoldi I, Yung AR, et al. Predicting psychosis: meta-analysis of transition outcomes in individuals at high clinical risk. Arch Gen Psychiatry. 2012;69:220–229. [DOI] [PubMed] [Google Scholar]
- 28. Haddock G, McCarron J, Tarrier N, Faragher EB. Scales to measure dimensions of hallucinations and delusions: the psychotic symptom rating scales (PSYRATS). Psychol Med. 1999;29:879–889. [DOI] [PubMed] [Google Scholar]
- 29. Kay SR, Fiszbein A, Opler LA. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull. 1987;13:261–276. [DOI] [PubMed] [Google Scholar]
- 30. Fusar-Poli P, Nelson B, Valmaggia L, Yung AR, McGuire PK. Comorbid depressive and anxiety disorders in 509 individuals with an at-risk mental state: impact on psychopathology and transition to psychosis. Schizophr Bull. 2014;40:120–131. doi:10.1093/schbul/sbs136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Kempton MJ, Bonoldi I, Valmaggia L, McGuire P, Fusar-Poli P. Speed of psychosis progression in people at ultra-high clinical risk: a complementary meta-analysis. JAMA Psychiatry. 2015;72:622–623. [DOI] [PubMed] [Google Scholar]