Schizophrenia is a serious and disabling mental illness that substantially impacts upon a patient’s capacity to achieve or maintain employment and relationships. In addition, schizophrenia has a huge societal effect, posing a challenge for families, caregivers, and healthcare systems.

Generally, early detection and intervention are associated with a better course of schizophrenia and an improved outcome. Most current guidelines recommend continued antipsychotic medication for 1–2-years following a first episode of schizophrenia [1–4]. During this period antipsychotic discontinuation is strongly associated with relapse [5], however, a minority of patients might not need medication to prevent relapse [5, 6]. On the other hand, withdrawing antipsychotic treatment according to the above-mentioned guidelines has been associated with a relapse rate of almost 80 % after 1 year post discontinuation and 96 % after 2 years [7]. Similarly, intermittent treatment, i.e., tapering medication once a patient is in remission and restarting it when early warning signs emerge, is less efficacious than continuous treatment, both in first-episode [8] and multiple-episode patients (e.g., [9]).

Hence, the effectiveness of antipsychotic drugs in reducing the risk of relapse is undisputed. In this context, a systematic review and meta-analysis of randomized, placebo-controlled trials revealed a significantly lower relapse rate during treatment with new-generation antipsychotics versus placebo [10].

Importantly, even a very short treatment interruption can increase the proportion of hospitalized patients. For example, Weiden and coworkers followed patients for 1 year and found that the risk of hospitalization doubled following a gap in medication therapy of 1–10 days, nearly tripled if medication was not used for up to 30 days, and quadrupled following a gap of more than 1 month [11].

Altogether, only about 50 % of schizophrenia patients can be expected to be fully adherent to treatment with medication, i.e., to take 80 % or more of the prescribed medication. The remaining patients are either partially adherent, taking 50 % or more of medication and either reducing the dose or failing to take medication from time to time, or may not follow prescription instructions at all [12, 13].

The degree of adherence is generally overestimated by psychiatrists, patients, and their relatives (e.g., [14, 15]). In addition, the attitude of professional careers toward the illness and the treatment plan are factors, which may influence patients’ attitudes toward the illness and medication and consequently adherence. For example, Rettenbacher and coworkers reported that only 35 % of nonmedical professionals, 46 % of psychiatric nurses, and 29 % of psychiatrists would be willing to take antipsychotic medication, if hypothetically suffering from schizophrenia. It is clearly extremely difficult, if not impossible, for staff to motivate patients to adhere to treatment recommendations, if they are not convinced about the need for medication themselves [16].

The majority of patients have adherence problems at some point [17]. A recent review of the literature reported rates of 15–40 % during the first month of treatment, and 50 % by treatment month 6 [18]. This proportion increases up to 75 % within 2 years [19]. Importantly, adherence rates are equally problematic for antipsychotic and nonpsychiatric medications [20, 21].

The clinical consequences of nonadherence to pharmacological treatment include an increased risk of relapse, rehospitalization, and antipsychotic treatment resistance [22, 23]. Further, nonadherence and partial adherence are associated with substance use, violence, arrests [24], suicide attempts [25], and poorer long-term functioning [24].

Clearly, patients’ reasons to adhere to a prescribed medication regimen and to remain on treatment are complex and variable. Interestingly, achieving symptom relief has been associated with poor adherence in first-episode patients, probably due to impaired insight into the need of maintenance treatment to prevent relapse [26]. On the other hand, as might be expected, nonadherence is more common when the patient perceives the benefits of medication to be suboptimal. In this context, it has been suggested that good adherence may lead to better response, which in turn leads to better adherence, while poor adherence may lead to poor response, which may further decrease adherence and increase the likelihood of treatment discontinuation [26].

In general, monotherapy results in better patient adherence than polypharmacy. Dosage regimen represents another pharmacological factor that may impact on adherence. Especially in elderly people neurocognitive impairment, which is commonly associated with schizophrenia, may represent a barrier to adhering to a complicated oral antipsychotic regimen [27]. A systematic review of the association between drug delivery systems and adherence to psychopharmacological treatment concluded that next to switching to oral slow-release formulations and depot injections a reduction in the frequency of administration, ideally a once-daily regimen, may improve adherence [28]. Similarly, Claxton et al. have shown that adherence of both mental health and nonmental health patients deteriorated with higher daily dosing frequency [29]. Interestingly, it has been reported that physicians are more likely to change or add medications in patients who are not fully adherent [30]. However, Bordenave-Gabriel et al. did not find an association between daily dose frequency and adherence [31], and in a recent European multi-center study a high daily dose frequency was associated with even better adherence [32]. Accordingly, patient preference for a specific regimen (even for more frequent dosing) may exert the dominating influence on adherence [33]. Furthermore, the route of treatment administration is discussed as having an influence on adherence. For example, positive attitudes toward liquid formulations or orally disintegrating tablets have been reported to indirectly facilitate adherence [34, 35]. The role of long-acting injectable antipsychotics is discussed in Psychological Issues in Improving Adherence and Alliance.

Compared to first-generation antipsychotics (FGA) new-generation agents (NGA) have less liability to cause EPS. It was hoped, therefore, that the increased use of NGAs would improve treatment adherence. Actually, some clinical trials have found a modest advantage for NGAs in this regard [62–66], however, a large number of studies failed to show such a difference [32, 67–73]. A recent study, for example, investigated the relationship between medication-related factors and adherence in schizophrenia outpatients in four European countries and found the attitudes toward antipsychotic medication to predict adherence, whereas the type and formulation of drugs as well as side effects were not meaningful in this context [32]. Hence, regular examinations of patients’ attitudes toward treatment and regular risk/benefit assessments may be of more relevance for medication adherence than the type of treatment.

Long-acting injectable (LAI) or depot antipsychotics are associated with a number of advantages as compared to oral compounds: when steady-state levels are achieved, a more controlled and constant input rate together with the circumvention of first-pass metabolism [74] is associated with a lower variability in the range of plasma concentrations [75], which in turn has been claimed to lower the risk of side effects (e.g., [76]). In addition, treatment with LAIs has been associated with early detection of relapse, improved relapse prevention, and reduced rehospitalization rates ([77–81], but see also [82]). Lastly, LAIs facilitate the differentiation between a lack of efficacy and poor adherence, which has important implications for subsequent treatment planning regarding the potential need to change the dosage or the medication. Furthermore, nonadherence is immediately obvious when the patient does not show up for the scheduled injection, thereby facilitating the opportunity for appropriate, early interventions including psychosocial treatment, which may result in improved long-term symptom control [22, 83]. However, treatment with LAIs clearly does not guarantee medication adherence. Weiden and coworkers, for example, found higher adherence rates in patients converted to LAI treatment during inpatient stay as compared to those who remained on oral medication at one month postdischarge, but no between-group differences in this regard at 6 and 12 months postdischarge [84]. Accordingly, factors such as medication beliefs and lack of insight can also reduce adherence to LAI therapies [85]. Nevertheless, in a number of studies treatment retention was significantly higher with injectable versus oral formulations (e.g., [78, 86, 87]).

Overall, it is important to see LAIs as a positive individual choice and not as a coercive method to solve difficulties with adherence.

Despite these advantages it is estimated that merely between one quarter and one third of people with schizophrenia are prescribed an LAI antipsychotic in the UK, a country in which LAIs are used more commonly than in many other countries [88]. On the part of the patients this might be due to fear that injections constrain their autonomy, are painful [89], or stigmatizing [90]. When asked to state a treatment preference (oral versus LAI), patients tend to favor their current formulation [91, 92]. Many clinicians, on the other hand, are hesitant to use LAIs despite regarding them to be associated with better adherence ([93, 94] but see also [95, 96]) and assume, that they are not useful in patients sufficiently adherent with oral medication, that they are associated with more side effects, that patients always prefer oral medication, and that they should not be used in first-episode patients [93, 97–99]. However, recent studies suggest that the use of LAI medication is feasible in first-episode psychosis. Emsley et al. for example, conducted a 2-year study and reported on symptomatic remission in 64 % of first-episode patients who were treated with LAI risperidone with 97 % maintaining this status until study completion. Notably, patients in remission received lower doses of antipsychotic medication, had fewer EPS, and a more favorable attitude toward medication [100]. Importantly, almost all patients, who chose to discontinue medication after the completion of the study eventually relapsed [101]. In a prospective study by Weiden et al. first-episode patients were randomly assigned to a recommendation of changing to LAI risperidone versus continuing on oral antipsychotic medication following clinical response to oral treatment. At 12 weeks, patients accepting LAI treatment (73 %) were significantly more likely to be adherent than patients staying on oral medication [102]. Moreover, in a naturalistic study partial or nonadherence occurred significantly more often in first-episode patients who were treated with oral risperidone as compared to those who were prescribed LAI risperidone [103]. However, further research is clearly warranted to address the question whether or not treatment with LAI antipsychotic medication reduces the risk of nonadherence.

As mentioned above, some psychiatrists [93] and psychiatric nurses [104] consider LAI antipsychotics to be associated with a greater risk of side effects than oral medication. To determine whether pharmacokinetic differences actually affect the tolerability of oral versus LAI formulations of an antipsychotic, double-blind, randomized studies comparing bioequivalent doses of the two drugs would be needed. As bioequivalence would have to be determined by comparing receptor occupancy and plasma levels, such studies are very difficult to perform. However, a number of studies have compared the pharmacokinetic and side effect profiles of antipsychotics in oral and LAI forms. In this context, a metaanalysis by Adams and coworkers revealed an equivalent rate of EPS (defined as the need for anticholinergic medication) and tardive dyskinesia for FGAs in oral and LAI formulations [105]. Among NGAs, risperidone LAI (RLAI) has been shown to cause less fluctuations in plasma levels and a lower mean steady-state Cmax concentration than bioequivalent doses of oral risperidone [106]. This could account for the greater decrease from baseline prolactin levels found in patients switching to treatment with RLAI as compared to those who continue treatment with oral risperidone [107, 108]. It is unclear, however, whether this laboratory finding is of clinical relevance. In patients treated with olanzapine, the pharmacokinetic differences between the oral and the LAI formulations seem to be irrelevant to the side effect profile [109]. However, when treating patients with olanzapine LAI (OLAI) one has to consider the risk of a post-injection syndrome, which presents with symptoms of an olanzapine overdose, occurs in less than 0.1% of injections, and appears to be unique to treatment with OLAI [110]. Therefore, after each injection of OLAI, patients have to be observed in a health care facility by an appropriately qualified person for a minimum period of 3 h. In addition, patients have to be accompanied home and should not drive or operate machinery on the day of injection. This clearly might have implications on a patient’s attitude towards treatment with OLAI and secondarily on adherence. In addition, these protective measures infer a strain on the care system, thereby restricting the feasibility of managing patients with OLAI. Specialized depot clinics may help to overcome these hurdles.

Injection-related adverse effects that can occur with any injectable antipsychotic include injection-site pain and a range of local injection-site complications (swelling, induration, etc.). Next to appropriate injection technique, rotating injection sites, avoiding excessive injection volumes, and increasing the injection interval represent potential strategies to minimize these risks.

A number of circumstances and parties are involved in the multifaceted issue of adherence in schizophrenia. Next to the establishment of a solid therapeutic alliance they include pharmacological strategies providing optimal efficacy along with minimal adverse effects, the prescription of a once-daily dosage regimen, and monotherapy, if possible. In addition, the use of long-acting injectable antipsychotics allows the most reliable assessment of adherence, as patients who regularly show up to receive their injection have assured adherence, whereas patients who do not, are easily identified to be nonadherent.