The core finding from this study of seriously ill community based schizophrenic patients is the predominance (70%) of off-label medication administration with the use of antipsychotic as a sole medication class (per FDA label) a relatively infrequent occurrence (25.5%). Despite emerging data suggesting relative therapeutic equivalence, SGA's were far more frequently administered than were FGAs (88 vs 21.5%); when FGA administration occurred it was overwhelmingly as an addition to SGAs: 65% of FGA administration was with concomitant SGAs. This study's sample was clinically well characterized; demographics are is reflective of highly symptomatic patients with schizophrenia in midlife. For example, there was a slightly higher male prevalence (59.5%); patients were largely never married (75%); a majority live in supervised dwelling (54.5%); and nearly half had had involuntary hospitalizations (47.5%), had been in jail (48.5%), and had made a suicide attempt (48%); mean age of onset was 19.9 years. These demographies place the study population into the mainstream of community based seriously mentally is patients with schizophrenia The relatively high prevalence of clozapine administration (18%) is consistent with the high severity of illness of our population as reflected by their behavioral ratings.. The relatively low prevalence of the Schizophrenia, Disorganized Type is likely due to the evaluation of patients through the “window” of medication treatment rather than in the drug free state. 79% of the population was overweight or obese, reflecting one of the numerous elements of enhanced health risk that most of these patients face. The most frequent combination of drug classes was a mood stabilizer added to an antipsychotic drug, followed by more than one antipsychotic, antipsychotic and antidepressant and antipsychotic and anti-anxiety combinations. The most commonly used APS were olanzapine (26%), and risperidone and quetiepine (20.5%) with frequency of clozapine administration at 18%.
Buchanan et al 
reported that 50% of 344 schizophrenic outpatients were treated with either antidepressants, anti-anxiety or mood stabilizers concomitant with antipsychotic drugs and that 17% of patients were treated with more than one adjunctive agent. Tapp and colleagues 
investigated the utilization of more than 1 antipsychotic and found in a survey of a diagnostically diverse group of schizophrenic outpatients that 13% received an FGA added to SGA, a comparable figure to the 13.5% frequency we observed for this combination. Baseline medication use of the 1,493 patients of the CATIE study 
revealed a high frequency of no antipsychotic medication (26%) and a low frequency of more than 1 antipsychotic (5%) and antipsychotic plus mood stabilizer (including lithium) (15%) in comparison to the prevalence of no antipsychotic (4.5%), patients treated with more than one antipsychotic (42.5%) and patients treated with antipsychotic plus mood stabilizer (44.5%) in a our cohort. Our cohort of community-based patients who were not participating in a prospective double-blind controlled study was likely considerably more ill than CATIE patients as reflected by mean total PANSS score: 111±16.5 in our cohort vs 75.7±17.6 for CATIE.
Interestingly, baseline CATIE antipsychotic plus antidepressant (31%) and antipsychotic plus anti-anxiety (18%) treatment combinations 31) were comparable to prevalence among our patients (38% and 14.5%, respectively).
The predominance of off-label drug combinations speaks to the overriding message of CATIE 
and CUtLASS 1 
: there are significant limitations in ineffectiveness of all antipsychotic drugs. There are, however, no clear standards or guidelines for the use of off-label treatments. Our multivariate model predicting use of antipsychotic drugs and concomitant mood stabilizers identified the diagnosis of schizoaffective schizophrenia, history of having hurt someone and high scores on the General Psychopathology subscale of the PANSS (which includes items such as “uncooperativeness,” “lack of judgment and insight,” “poor impulse control,”) as predictors. This suggests combined antipsychotic and mood stabilizers are used in patients with aggressive elements to their behavior. In contrast, greater number of hospitalizations and high Montgomery-Asberg depression ratings were predictors of concomitant antidepressant use, suggesting this approach in depressed schizophrenics 
with high risk of relapse. We are unaware of data elsewhere related to clinical predictors of off label drug administration.
There are two critical elements to off-label prescribing practices related to the Food and Drug Administration: a drug approved for marketing may be labeled, promoted and advertised by the manufacturer for only those uses for which the drug's safety and effectiveness have been established by the FDA.
Industry practices regarding promotion of uses not included in the drug label have become increasingly scrutinized, as exemplified by the attention and penalty to market practices that encourage off-label use of the anticonvulsant, gabapentin.
The FDA has recently proposed new guidelines that enable sponsors to distribute publications about unapproved uses of approved drugs and advices. 
Of serious concern, however, is that the selective use of peer-reviewed literature may not be able to satisfactorily ensure the quality of off-label promotion 
, contributing to the problematic oversight of industry's promotional efforts. In contrast to industry whose “behavior” in the marketplace is at least theoretically closely scrutinized, the clinician has considerable flexibility: if a product has been FDA approved, a physician may choose to prescribe it for uses or in treatment approaches or patient populations other than the approved indication
. It is the responsibility of the manufacturer to gain FDA approval for adding new uses to the product label. It is hardly surprising that a company may be hesitant or even resistant to invest the resources and entertain the risk of unfavorable results involved in FDA review of a new indication, given the multibillion dollar revenues for medications whose off label use in schizophrenia is described in this report Moreover, the impact of off label use in schizophrenia is particularly great on the public sector as schizophrenic patients' care is largely supported by Medicaid and to a lesser degree Medicaire. In a recent Wall Street Journal/Harris poll 
, the public appears evenly divided on whether physicians should (45%) or should not (46%) be allowed to prescribe medications for off-label uses; in contrast, a majority (62%) of respondents believe that pharmaceutical companies should not be allowed to encourage off-label use.
The scientific merits underlying the use of these off-label drug class combinations are variable; although it is an area where clinicians play a major role in the development of drug treatment
Radley et al 
used the DRUGDEX 
system, a highly recognized scientific documentation resource, to categorize off-label uses as having strong scientific support, limited scientific support or no scientific support. Their findings that 96% of psychiatric off-label uses have limited or no support might well be questioned by the psychiatric research community. The need for systematic evaluation of treatment efficacy of drug class combinations is clearly needed.
In summary, there is a predominance of off-label prescription use in the treatment of seriously ill patients with schizophrenia in the community. It appears that the real world pharmacotherapy of schizophrenia has developed its own established practice that may have outstripped current data support. The economic implications of off label use in schizophrenia for public sector payers as well as for the pharmaceutical industry are substantial. The independent research community could make an important contribution by supporting a program of systematic evaluation. What might such an undertaking look like from the perspective of clinical trial design? One clear and logical approach is to study the superiority, on some primary endpoint (e.g. total PANSS score) when the drug in question is added to an antipsychotic in comparison with antipsychotic monotherapy. In light of the very high drop out rate of the ambitious CATIE study, a design that enabled a high rate of subject completion (CATIE completion rate: 26%) would certainly be necessary for the study to have the necessary impact. It's unlikely that such work will stem from industry sponsored initiatives.