- Top of page
In the last 15 years, a number of studies have supported the view that the earlier detection and care of psychosis can lead to a better outcome.[1, 2] However, most of these studies were conducted in research settings; thus, the actual effectiveness of early intervention for psychosis remains unclear. One of the issues that such studies have raised is that the diagnostic criteria or primary outcomes focus mainly on the attenuated positive psychotic symptoms.
Attenuated psychotic symptoms or psychotic-like experiences have been commonly found in the general population, and these symptoms or experiences may not necessarily be associated with distress or help-seeking behaviour.[4-6] In a previous study comparing help-seeking patients with the general population, the authors reported that psychosis-like experiences do not significantly contribute to help-seeking behaviour. Attenuated positive symptoms may not always confer subjective difficulties or sufferings; therefore, the current interventions to reduce risk which are focused on the attenuated positive symptoms may not be truly capable of meeting the needs of individuals meeting at-risk criteria.
To date, longitudinal studies on the outcomes of individuals at risk for psychosis have underlined the considerably high rates of remission and the low rates of transition to psychosis.[9, 10] Given that the criteria for remission and transition are based on the attenuated psychotic symptoms, however, it would be doubtful whether or not these outcomes reflect the actual changes in subjective difficulties of individuals at risk for psychosis. Indeed, a large longitudinal study, the North American Prodrome Longitudinal Study (NAPLS), revealed that most individuals who met the at-risk criteria but did not convert to psychosis continued to suffer from lower levels of functioning or disabilities. Additionally, an approach focused predominantly on the low rate of transition to psychotic disorder can obscure individual treatment effects. Subgroups of participants may respond to individual treatments particularly well or particularly poorly as a result of the participants' characteristics or baseline symptom patterns. A recent report on a randomized controlled trial examining the effect of various therapies on young people with a high risk for psychosis concluded that the interventions were equally effective or ineffective. Thus, the effectiveness of interventions for early psychosis should be clarified in clinical settings, regardless of the transition to full-blown psychosis.
We hypothesized that the current strategy, which is focused on the attenuated positive symptoms, cannot sufficiently ameliorate the subjective difficulties of individuals at risk for psychosis, such as their subjective quality of life (QOL), role/social functioning, interpersonal relationships and subjective well-being. We also assumed that some patients would continue to receive treatment because their symptoms had not been relieved.
The aims of this study were: (i) to describe the actual 1-year outcome of individuals with a high risk of psychosis based on comprehensive assessments including subjective QOL, role/social functioning, interpersonal relationships, insight into illness and subjective well-being; and (ii) to clarify the characteristics of patients who continue to receive treatment for over 1 year so as to explore the factors that may lead to a poor outcome, even without a transition to psychosis.
- Top of page
At baseline, 46 treatment-seeking patients who had been clinically diagnosed as having clinical high risk of psychosis agreed to participate in the study and to be assessed. The demographic characteristics of the sample at baseline are presented in Table 1.
Table 1. Demographic characteristics of the sample at baseline (n = 46)
|Past treatment history||29||63.0|
|Family history (any mental illness)||18||39.1|
|Duration of illness, weeks||26.0||24.0|
|SOPS|| || |
|SFS|| || |
|SWNS|| || |
|WHO-QOL26|| || |
|SUMD, current disorder|
|Item 1-3 (global insight) awareness||2.3||0.9|
|Item 4-10 (symptom items) awareness||1.5||0.5|
|Item 4-10 (symptom items) attribution||3.0||0.9|
At the 1-year follow-up point, 27 participants (59%) completed the reassessment. Table 2 shows the sample characteristics of these 27 patients and the patients who withdrew from the study, indicating that the withdrawn patients were younger and had a shorter duration of illness, less negative/general symptoms and a higher QOL.
Table 2. Comparisons at baseline between the followed-up patients and the withdrawn patients
| ||Followed-up (n = 27)||Withdrawn (n = 19)||Chi-square||P|
|n ||%||n ||%|| || |
|Past treatment history||21||77.8||8||42.1||0.01||0.90|
|Family history (any mental illness)||10||37.0||8||42.1||0.12||0.77|
|Duration of illness, weeks||30.7||24.5||19.2||21.9||−2.08||0.04*|
|SOPS|| || || || || || |
|SUMD, current disorder|
|Item 1-3 (global insight) awareness||2.3||0.9||2.4||1.0||−0.45||0.65|
|Item 4-10 (symptom items) awareness||1.6||0.5||1.3||0.3||−2.04||0.04*|
|Item 4-10 (symptom items) attribution||3.0||0.9||3.1||1.0||−0.02||0.99|
During the follow-up period, three patients, or 12% of the followed sample, converted to psychosis: two were diagnosed as having schizophrenia and one was diagnosed as having a schizoaffective disorder. According to the criteria mentioned above, 14 patients were defined as ‘improved’ (in both the SOPS positive and negative symptoms), but 13 patients, including the 3 psychotic cases, were defined as ‘not improved’ (in either the SOPS positive or negative symptoms). Detailed comparisons of these two groups are shown in Table 3, suggesting that although few differences in the clinical variables were found between the two groups at baseline, all the patients in the ‘not improved’ group had past treatment histories and had fewer family members with mental health illness.
Table 3. Comparisons at baseline between the ‘improved’ group and the ‘not improved’ group
| ||Improved (n = 14)||Not improved (n = 13)||Chi-square||P|
|Past treatment history||8||57.1||13||100.0||7.16||0.02*|
|Family history (any mental illness)||8||57.1||2||15.3||5.04||0.04*|
|Duration of illness, weeks||34.4||26.0||26.7||23.3||−0.95||0.34|
|SOPS|| || || || || || |
|SUMD, current disorder|| || || || || || |
|Item 1-3 (global insight) awareness||2.5||0.9||2.1||0.8||−1.42||0.16|
|Item 4-10 (symptom items) awareness||1.5||0.5||1.6||0.5||−0.65||0.52|
|Item 4-10 (symptom items) attribution||2.9||1.1||3.1||0.8||−0.65||0.52|
Table 4 shows that ‘not improved ’group demonstrated a decline of the SWNS total score and the SUMD sub-score (awareness of symptoms) over time, even after adjusting for age, duration of illness and baseline scores. Twenty-one (78% of the followed) patients had received antipsychotic medication at the follow-up point (aripiprazole: n = 13; quetiapine: n = 5; perospirone: n = 2; risperidone: n = 1), whereas only six patients (22%) were administered antipsychotic treatment at baseline (quetiapine: n = 2; risperidone: n = 2; aripiprazole: n = 1; perospirone: n = 1) (Table 3).
Table 4. ANOVA for comparing clinical outcomes between the ‘improved’ group and the ‘not improved’ group
| ||Score difference (T2-T1; mean ± SD)||Non-adjusted||Adjusted†|
|SWNS total||17.5 ± 17.3||2.4 ± 17.3||4.896||0.037*||5.125||0.034*|
|SFS total||13.7 ± 21.7||4.6 ± 15.0||1.509||0.231||1.575||0.223|
|GAF||19.3 ± 11.6||12.7 ± 17.4||1.363||0.254||3.058||0.094|
|WHO-QOL total||3.7 ± 3.8||1.2 ± 3.9||2.855||0.104||1.024||0.323|
|SUMD global insight||−0.3 ± 1.0||0.2 ± 0.9||0.484||0.494||0.005||0.947|
|SUMD symptom awareness||0.9 ± 0.9||−0.1 ± 0.7||8.632||0.008**||8.435||0.009**|
|SUMD symptom attribution||0.5 ± 1.0||0.1 ± 0.3||0.645||0.432||0.647||0.432|
Multiple linear regression analysis was used to explore variables at baseline that can predict poorer outcome at the follow-up point (Table 5). Results suggest that less severe negative symptoms, more severe general symptoms, or lower subjective well-being at baseline could significantly predict poorer outcome after 1 year.
Table 5. Multiple linear regression analysis for exploring variables that can predict poor outcome at the follow-up point
|Negative symptoms at T1||−0.060||0.020||−0.525||−2.907||0.008|
|General symptoms at T1||0.174||0.039||0.915||4.454||<0.001|
|SWNS total score at T1||0.024||0.009||0.510||2.677||0.014|
- Top of page
Our findings are of some clinical relevance when treating help-seeking individuals with the features of early psychosis. The current naturalistic study revealed that quite a few patients (48%) showed little improvement in both their positive/negative symptoms and subjective well-being after having received intervention for over 1 year, regardless of transition to full-blown psychosis. Additionally, nearly half of the entire sample (41%) dropped out of the study within 1 year for any reason. These results suggest that the current early interventions cannot truly meet the subjective needs of individuals at risk for psychosis.
One explanation for the unmet needs among the at-risk patients might be that the early interventions for psychosis in clinical settings tended to favour antipsychotic medication, as seen in the present study. We found that about 80% of the patients who were followed up had received antipsychotic medication at the follow-up point. Although such antipsychotic medication would be generally administered to reduce risks that are focused on the attenuated positive symptoms, the results indicated that poorer outcome could not be significantly predicted by severity of positive symptoms at baseline but less severe negative symptoms, more severe general symptoms, and lower subjective well-being at baseline. This suggests that other symptoms than positive symptoms might be a key to patients' subjective difficulties in their daily lives, possibly shedding new light on early intervention strategies for psychosis; for example, a targeted intervention for affective symptoms might be more effective with regard to the subjective response than interventions for positive symptoms.
In addition, to make matters worse, the off-label use of antipsychotics for psychosis prodrome has presented some ethical issues associated with unexpected adverse effects, social stigmatization and low self-esteem. Given that poor adherence to the initial treatment may hinder an adequate intervention, ethical issues regarding pharmacological intervention during the earliest stage of psychosis cannot be ignored. However, recent clinical research has revealed that not a few clinicians in the community have administered pharmacological interventions, including antipsychotics, to individuals who have attenuated psychotic symptoms but do not meet the criteria for psychosis. A naturalistic study from the Recognition and Prevention program showed that individuals presenting with more severe (but non-psychotic) attenuated positive symptoms were nearly all treated with antipsychotics, often in combination with other agents. The data from the NAPLS demonstrated that 60% of the clinical high-risk sample had a lifetime history of receiving psychotropic medication prior to their entry in the research program. Also, anonymous surveys in Japan and Singapore have indicated that most psychiatrists in the community would treat prepsychotic patients with active management, including antipsychotic medication.[23, 24] Generally, most clinical psychiatrists in the community are likely to overestimate the use of pharmacological intervention, including antipsychotics, for individuals who have attenuated (but non-psychotic) psychotic symptoms. However, as a number of medication-free studies have found, antipsychotic medication does not seem to be an essential component of effective treatment for psychosis, even in patients with established illnesses.
The high dropout rate in the study (41%) may be partially due to this strategy for intervention that was focused on attenuated positive symptoms. The patients who withdrew within the 1-year follow-up period were younger and had shorter duration of illness, less severe negative symptoms/general symptoms, better awareness of symptoms, and higher subjective QOL at baseline than the patients who were followed up. Although the reasons for dropping out are needed to be explored, it is noteworthy that although the withdrawn patients had better clinical characteristics at baseline, there were no significant differences in positive symptom at baseline between the withdrawn patients and the followed patients. This result suggests that adherence to treatment in individuals with clinical high risk of psychosis does not depend on the extent to which interventions are based on the target for reducing positive symptoms.
Other clinical variables may also have some impacts on treatment outcome. Patients in the ‘not improved’ group had past treatment histories and had fewer family members with mental health illness. There are two potential interpretations for this finding. First, it may be that those with family experience of psychiatric illness tended to have effective care or support during the earlier stage of illness. Although previous studies have failed to confirm that family history of psychiatric illness was positively associated with a shorter duration of untreated psychosis,[25-27] families with previous experience of mental health illness may facilitate earlier help seeking through the enhancement of knowledge about potential symptoms and their significance. Second, patients in the ‘not improved’ group may be treatment resistant. These patients would continue to receive treatment because their symptoms had not been relieved, as we hypothesized, partly because the current early interventions were not effective for this type of patients. Another explanation for considerable rate of having past treatment history is the preponderance of women in the present study sample. Several studies showed that women in general are more likely to have a past history of any psychiatric disorder.[28, 29] Given that gender differences may influence the course of illness, our results would be skewed by the predominance of women in this sample.
Our data further suggest that negative symptoms do not appear to have an impact on both clinical outcomes and treatment adherence. Less severe negative symptoms at baseline were found to be associated with withdrawal from treatment and to predict significantly poorer outcomes, contrary to previous findings.[30, 31] These findings are also contrary to our previous expectation that severe negative symptoms would be associated with withdrawal from treatment and poorer outcomes. Rather, it appears that general symptoms play a key role more than negative symptoms for both clinical outcomes and treatment adherence. Whereas less severe general symptoms at baseline were found to be associated with withdrawal from treatment, more severe general symptoms at baseline predict poorer clinical outcomes after 1 year. General symptoms include sleep disturbance, dysphoric mood, motor disturbance and impaired tolerance to normal stress. These symptoms may be directly linked to difficulties in daily living, in other words, subjective difficulties. Therefore, fluctuation of general symptoms should be carefully evaluated as a measure of effectiveness in the treatment.
The present study had some methodological weaknesses. First, an evaluation of the extent of the patients' needs is needed to clarify the relationship between subjective difficulties and help-seeking behaviour. Subjective difficulties would be hard to be evaluated precisely by the objective ratings and thus further development of objective ratings on subjective wellness/difficulties should be needed. Second, a considerable attrition rate was also observed in the current study, as in most prospective studies, but the reason for the high rate of patients lost to attrition remains unclear. Third, the present sample was skewed by both this high attrition rate and high rates of previous treatment with relatively long duration of ‘being well’. Finally, the small number of subjects in this study may certainly limit the generalizability of the findings. A larger sample with a longer period of observation is needed.
Despite these limitations, our findings have important clinical implications. A notable number of patients had a poor outcome with symptomatic deterioration, providing a rational for early intervention for psychosis. However, the current strategy for reducing the risk of psychosis, which is focused on the attenuated positive symptoms, should be reappraised. Further comprehensive longitudinal studies are needed to develop truly needs-based interventions for these at-risk patients.