Early intervention in psychosis (EIP) is not a recent idea and delaying/preventing young people from psychosis has always been an aspiration.

Indeed, throughout during the twentieth century the prospect and practicality of identifying subjects in the prodromal phase of schizophrenia was debated. The prodromal symptoms and signs were pointed out, but they were extremely non-specific and not useful from a clinical point of view. For example, E. Bleuler (1911) [12] maintained that it was not possible to talk about prodromes, because nearly every symptom can be a forerunner of schizophrenia and at the beginning, when the symptom appears, it is impossible to predict accurately the future evolution of the disease. Nevertheless, Bleuler himself, and many other psychiatrists of the past century, thought early intervention offered the only possibility to improve the outcome of psychotic disorders, even though individual symptoms were not specific enough.

Many other authors (i.e. [1, 13–19]) agreed that psychotic onset should be considered a culmination of an evolving illness. In this chorus of unanimous voices, two eminent psychiatrists struck a discordant note. Schneider and Jaspers in fact, thought that schizophrenia was independent of a premorbid personality, and that it was impossible to detect a pattern of prodromal symptoms. This was the prevailing position during the 1960s, also due to a widespread transition taking place from qualitative to quantitative studies in psychiatry. With the attempt to align psychiatry to other medical disciplines [20] and the consequent development of DSM-III [21], American psychiatry progressively lost interest in conditions with minimal symptomatic expressions and prodromal stages. The limited space reserved for prodromal symptoms in schizophrenia in the DSM-III, disappears completely in the DSM-IV [22], where the possibility of prodromal phases subsists only for cluster A personality disorders.

Classic psychopathologists’ research was eclipsed by the pressing need for a valid and widely shared diagnostic system, which can unify and enable comprehension of professionals worldwide. The return to a Kraepelinian model, which requires clear and evident psychotic signs and symptoms exhibited for an extended period before a formal diagnosis, closed the debate on prodromal forms of schizophrenia.

In the 1960s, two research areas developed, both proving to be extremely useful for an evaluation of prodromal phase and schizophrenia prevention: the work of James Chapman et al. [23] on psychotic-like experiences (PLEs) in the United States and the research of Gerd Huber [16] on basic symptoms in Germany.

Regarding PLEs, Chapman [23] and other colleagues like Gillies [24], Meehl [25], etc. hypothesised a continuum in the disease, which evolves progressively from the initial phases to frank psychosis. Huber [16], instead, supported the idea of the existence of ‘basic’ symptoms, which are mild subclinical and subjective symptoms involving instincts, affect, stress tolerance, thought, language, perception and motility.

What is new in the last 20 years is the attempt to find subjects at risk and to treat them at an earlier stage of the disorder. This paradigm shift entails a new approach to mental illness, particularly schizophrenia. At the beginning of the 1990s, as we highlighted in the previous paragraph, the Kraepelinian model, which had dominated during the century and established the basis for the new psychiatric nosology (DSM), was contested.

The neo-Kraepelinian model fallacy, McGorry says, is to posit a double bond between the disorder and its effects, the symptoms and signs and the diagnosis and prognosis, whereas psychosis psychopathology, in fact, is more fluid [26]. The connection between signs and symptoms, diagnosis and prognosis is much weaker and this opens the possibility for a secondary prevention of psychosis that is considered impossible or implausible in the neo-Kraepelinian model.

Within this paradigm shift, in the last decade researchers have tried to identify subjects ‘at risk’ and to intervene early in psychotic disorders [27].

Indeed, the poor predictability of prodromal symptoms has led to the application of a technique, which aims to refine diagnostic capacities and is already widely used in general medicine: the so-called ‘close-in’ strategy. This strategy consists in identifying the co-presence of many prodromal symptoms, a state–trait model that permits us to better identify cases at risk with a subsequent reduction of false positives. This type of evaluation method, first theorised and applied in Australia and the United Kingdom [27, 28], has also been utilised by many other research groups [29–32] in order to be able to offer a quick and specific response to the first signs or symptoms of the disease and to delay/prevent the psychotic onset [33]. The problem of early symptoms of psychosis is also very important because most young people who need help do not turn spontaneously to a CMHS [34].

Following and in coherence with the paradigm shift described above, an EIS specialised in management of FEP and UHR began to develop in the 1990s. These services developed roughly in parallel in England [3] and Australia [4], and spread to many other European nations, the United States of America, Canada, Singapore, Hong Kong and Japan. This model shift has become increasingly popular and has been applied in a number of countries. The main reason for this ‘success’ is the perception of the need for EIS instead of CMHS, since CMHSs were not structured to effectively engage and respond to the special needs of young individuals at their psychotic onset.

In the following paragraph we will describe the most important targets on which early intervention focuses and how EIS try to address their interventions specifically to these targets.

If we detect early signs and symptoms we may hope to delay, or even totally prevent, the development of a full-blown psychosis. This surely represents the most alluring and promising, although controversial, aspect of the field. Intervening in the prodromal phase means intercepting the subject in the period in which signs and symptoms of distress are present, without a real experience of psychotic rupture having taken place. However, since prodromal symptoms are nonspecific, the diagnosis is particularly difficult and the number of false positives is high and remains high [36], even though sensitive criteria have been developed to measure a condition at high risk of transition to psychotic onset. In the European Prediction of Psychosis Study (EPOS), a European, multicenter and prospective study, the combination of attenuated psychotic symptoms (APS) and cognitive basic symptoms (COGDIS) showed the best indication of risk to develop a psychosis in the following 18 months, with 48% of evolution rate [37]. Nevertheless, this means that in half of the at-risk subjects, psychosis does not emerge in the following 18 months. A recent meta-analysis [36] shows that the risk to develop a psychotic episode in a cohort of UHR subjects is 18% at 6 months, 22% at 1 year, 29% at 2 years and 32% at 3 years.

Guidelines for interventions in prodromal phases [35] support evaluating and taking charge of individuals with at-risk mental states, who ask for help. They should be followed over time. They should be offered psychological support together with treatment for specific symptoms such as depression, anxiety and substance abuse and should be assisted with eventual interpersonal, family and professional difficulties. Psychoeducational interventions are useful to develop coping strategies regarding subthreshold psychotic symptoms, together with supportive interventions and family psychoeducation. Treatment with antipsychosis medication at low doses is not indicated with the exception of people at high risk of suicide, rapid psychosocial deterioration or aggressive behaviours. When indicated and effective in the first 6 weeks, with the individual’s consent, pharmacological treatment can be continued for 6 months to 2 years. A gradual dosage reduction up to the minimum effective dose is recommended or, in some cases, an interruption of pharmacological treatment after 2 years in a consenting and responsive patient. When an antipsychosis medication is not effective, the administration of a different, second generation one is recommended. When the subject does not ask for help spontaneously, it is possible to get in touch with family and friends.

Evidence that these interventions are effective in reducing transition to frank psychosis is preliminary for now [38]. Further studies are needed to document the effectiveness of these interventions on the course of the illness.

Indeed, the risk/benefit ratio of early intervention in UHR is still unclear and the hypothesis of introducing a ‘psychosis risk syndrome’ (PRS) in the next DSM-V [39] stimulated a wide debate. At the present time, many researchers believe that introducing a PRS diagnosis in the DSM-V would be premature, mainly because of the significant heterogeneity of UHR subjects, their different clinical needs and prognosis [36]. It would be premature also since there is a lack of research confirming the reliability of standardised procedures for diagnosis, and this does not permit diagnostic validity. Biological and neuropsychological parameters should also be taken into consideration as a support to diagnosis. Only once research is able to provide answers to these questions will it be possible to include the diagnosis of PRS in DSM-V [37]. On the other hand, some authors support the inclusion of PRS diagnosis in DSM-V arguing that individuals with UHR are already ill and with poor functioning and are at high risk to worsen [40]. Moreover, DSM-IV diagnostic criteria are not specific for this condition [41]. There is broad agreement though, on the important role of research over the coming years to clarify the matter and to finally identify and treat subjects at-risk of psychosis.

Despite these different positions regarding UHR criteria and PRS, there is sufficient agreement about the desirability of early intervention for UHR. The main issue regards how this early intervention should be provided. Since CMHSs are not focused on prevention, EIS for UHR are more effective in providing the specific multidisciplinary treatments suggested by international guidelines.

Nevertheless, several authors [10, 11] suggest that the EIS model can pathologise normal developmental processes and lead to over-medicalization. This is not a minor question since it is not ethically acceptable to run the risk of treating people who are not really ill. Moreover, we know that PLEs are present in a large percentage of the population [42].

To avoid treating false positives, two main precautions have been developed:

1. People reach EIS for UHR by seeking help. Services do not screen adolescents in school and do not compel them to attend for treatment.
   
2. UHR criteria have been introduced to avoid this risk. Indeed, these criteria tend to identify people with a higher risk to develop psychosis and with a good response to treatment. They are identified by the presence of subthreshold psychotic symptoms.

In conclusion, EIS is offered only to the help-seeking, distressed individuals whose psychotic symptoms would be considered too mild to be considered ‘properly’ psychotic.

Timely detection of psychotic onset allows to reduce the duration of untreated psychosis (DUP), that is the time between the onset of positive psychotic symptoms and the initiation of appropriate treatment. Much evidence supports the view that the longer the time in which a person lives in a situation of frank psychosis without being treated, the worse is the progression and prognosis of the illness. DUP reduction is very important for short- and medium-term outcome [43]. A short DUP is linked to a better response to antipsychosis medication treatment [44] and better social and professional functioning [45ߝ47]. As a consequence of the reduction in DUP, individuals have a better response to treatment since they present less severe positive and negative symptoms and better social functioning than cases with long DUP [48].

The relationship between treatment delay and poor outcome has been clearly established [43, 44]. As we will highlight in this chapter, CMHSs are more often related to a consistent treatment delay as compared to EIS [49], and this supports the idea that specialised pathways to care can reduce DUP and consequently improve outcome.

The critical period (CP) is the period of time between 3 and 5 years that follows psychotic onset [27]. The years immediately following the psychotic onset are considered crucial in terms of prognosis [27, 50, 51], and what occurs in this phase can determine the following progression and prognosis of the disorder. In fact, during CP a dramatic progression towards a massive deterioration in social and cognitive skills occurs. This deterioration, once established, reduces the effectiveness of interventions. Indeed, the course of psychosis is characterised by a rapid deterioration in the first period, following which it reaches a plateau where it stabilises without any further worsening or improvement. It is necessary, therefore, to intervene as soon as possible to reduce or prevent psychosocial deterioration. As we indicated above, the crucial period is 3 to 5 years after the first psychotic episode [35, 52] and all international guidelines suggest continuous, intensive and assertive interventions during this phase. It is very important to treat the patient on time and continuously, since after the psychotic onset there is a high risk of relapse, severe disability and increased suicide risk. The entire family should be supported by psychoeducational group therapy. Psychological and psychosocial treatments should be the key elements during the CP to reduce positive and negative symptoms, to face eventual comorbidities and to promote social recovery. Treatment with antipsychosis medications must be accurately monitored to achieve the minimum therapeutic dose in order to reduce side effects that prevent recovery (weight gain, sexual dysfunctions and sedation) [35].

If, on the one hand, EISs are specifically developed to provide intensive and assertive intervention focalised in the CP, on the other hand some criticism of this approach has been voiced, since some authors argue that only individuals who fulfil the DSM-IV B (marked social deterioration) and C (duration of illness of at least 6 months) criteria for schizophrenia should be treated with this intensive model [10]. Although the position regarding treatment should be as conservative as possible, we should be aware that those criteria encourage delayed intervention which is harmful to health, especially in a population like this one, at high risk of death by suicide during the CP.

A larger randomised clinical trial, called OPUS [58], was run in Denmark on 547 individuals with first-episode psychosis. Trial design considered clinical evaluation of the effects (level of reduction of negative and positive symptoms) of specialised care based on the early intervention model in psychotic onset (assertive community Treatment, family care and social skills training), compared with standard care from territory psychiatry.

In the present historical period, in which resources are limited and it is imperative to stick to the budget, it is fundamental to evaluate the economic burden of EIS and compare it with those of the services that are already available. This evaluation should consider all cost variables derived from the illness, which means direct costs (directly attributable to medical care), indirect costs and, where possible, the so-called intangible or psychosocial costs (see [66, 67]).

In this direction some studies have compared specialised care for early psychosis using standard services by cost-effectiveness analysis. McCrone et al. [68] recently published a study on costs and cost-effectiveness regarding the LEO trial. Clinical variables and direct costs were measured relative to care management of individuals at accrual and then after 6 and 18 months. At the same time, information on quality of life and work re-engagement was gathered to evaluate cost-benefit. Relative to total direct costs, no statistically significant difference appeared between individuals in the specialised care branch as compared with those in standard care branch (£11.685 vs. £14.062). Cost-benefit analysis was clearly in favour of specialised care, showing a better clinical outcome with equal costs. In Australia, Mihalopoulos et al. [69] verified a reduction of direct costs in areas where an EIS was instituted, due to reduction of inpatient costs, especially in the medium-long term, after the service is fully operational [70].

Another interesting study highlights a reduction of direct costs and an improvement of the cost-benefit ratio after the institution of EIS for psychosis. Valmaggia et al. [9], in fact, analyzed the costs of their service allocated in south London, showing that costs do not differ from standard care in the first 12 months, but at 18 months a significant reduction begins to appear. Robust evidence of this cost reduction comes from a recent study [71], where it is estimated that over a 3-year period an EIS for FEP saves the NHS £15.862 per person with first-episode psychosis when compared to standard services. That is a potential £119 million saving for the NHS.

In summary, studies published up until now highlight a lack of increase, if not a downright reduction, of direct costs, with an improvement in the cost-benefit ratio.