The committee reviewed the data available on the clinical and cost effectiveness of aripiprazole, having considered evidence on the nature of schizophrenia in people aged 15 to 17 years and the value placed on the benefits of aripiprazole by people with the condition, those who represent them, and clinical specialists. It also took into account the effective use of NHS resources.
The committee discussed current standard clinical management of schizophrenia in adolescents. It heard from clinical specialists and patient experts that antipsychotics are prescribed only after a psychological assessment and a discussion with the person with schizophrenia together with their family or carer. The choice of treatment is negotiated with the person and depends on a number of factors, including adverse events associated with the treatment, previous treatments the person has received and their responses to them, and adverse events experienced while on those treatments. Adolescents with schizophrenia are usually treated with atypical antipsychotics at a low dose and are closely monitored.
The clinical specialists noted that the main aim of treatment is to maximise the control of schizophrenia and minimise the adverse events that are the most troublesome for each individual. The committee heard from the patient experts that effective control of schizophrenia with aripiprazole would allow adolescents to return to normal functioning in terms of work or schooling. The committee understood from the clinical specialists that no single atypical antipsychotic drug is considered to be more clinically effective than the others. Risperidone is the most widely used first-line atypical antipsychotic in UK clinical practice because clinicians have extensive experience of using it to treat schizophrenia, and often achieve control with low doses and without troublesome adverse events. The clinical specialists stated that when an atypical antipsychotic medication is prescribed, control of schizophrenia and adverse events is assessed over a period of 6 weeks or more and an alternative atypical antipsychotic can be considered if the first antipsychotic proves unsatisfactory. Other atypical antipsychotics such as aripiprazole, olanzapine, quetiapine or amisulpride may be used if control of schizophrenia is not achieved with risperidone. The clinical specialists also explained that clozapine is sometimes prescribed; however, because it needs careful monitoring for particular side effects, it is prescribed as rescue therapy only if the schizophrenia is refractory to at least 3 other antipsychotic treatments. The committee noted that some of the atypical antipsychotics described by the clinical specialists do not have a marketing authorisation for the treatment of schizophrenia in adolescents, but acknowledged that specific licensing in adolescents is not a prerequisite to prescribing licensed adult medicines, particularly if there is widespread experience of their use. The committee agreed with the clinical specialists that it is important for adolescents with schizophrenia to have a range of treatment options before considering rescue therapy with clozapine, and therefore considered that aripiprazole may be a suitable treatment option for people aged 15 to 17 years with schizophrenia.
The committee heard from the clinical specialists and patient experts that there are a number of dose-related adverse events associated with atypical antipsychotic treatments, including weight gain, hyperprolactinaemia and sexual dysfunction, aggression, and akathisia and extrapyramidal symptoms. Adolescents are often less tolerant of adverse events than adults, leading to problems with adherence to medication. The committee heard from the clinical specialists that some treatments are more likely to be associated with particular adverse events than others: olanzapine is more likely associated with weight gain, risperidone and amisulpride are more likely associated with hyperprolactinaemia, and aripiprazole is more likely associated with akathisia and a subjective feeling of aggression (for which benzodiazepine co-treatment may be used). The clinical specialists stated that these adverse events are dose related and therefore it is preferable to start prescribing any atypical antipsychotic at a low dose. The committee accepted that all atypical antipsychotics are associated with adverse events and that accounts from the clinical specialists on the use of aripiprazole suggest that it may be as safe and well tolerated as the other treatments.
The committee noted that the clinical effectiveness evidence presented in the manufacturer's submission was derived mainly from 1 RCT (study 31-03-239) that studied treatment with aripiprazole at 2 different doses compared with placebo, with supporting data on adverse events from 2 open-label single-arm extension studies. The committee noted that the 31-03-239 study was placebo controlled and did not provide a head-to-head comparison of aripiprazole with any other atypical antipsychotics.
The committee noted that the 31-03-239 study showed a reduction in total PANSS score (that is, an improvement in symptoms) at week 6 in all 3 study arms. Statistically significant differences in the degree of improvement were observed in the aripiprazole groups compared with the placebo group (p=0.0414 and p=0.0061 for the 10 mg and 30 mg doses versus placebo). The committee heard from the clinical specialists that the PANSS score is a well-recognised tool used in clinical trials for the measurement of positive, negative and general psychopathology symptoms in schizophrenia. However, the results are often difficult to relate to UK clinical practice as the tool is not routinely used by clinicians. The committee accepted that the PANSS score is a valid tool for the measurement of positive, negative and general psychopathology symptoms and that evidence from the 31-03-239 study demonstrates a reduction in schizophrenic symptoms in the aripiprazole groups.
The committee was aware that the manufacturer's original submission included a very limited evidence base. However, the manufacturer's additional analyses provided some evidence for each of the atypical antipsychotics (risperidone, quetiapine and olanzapine) routinely used in UK clinical practice. The committee noted that the trials for olanzapine, quetiapine and, most notably, risperidone showed greater relative risks in PANSS positive and negative scores in their treatment arms compared with their placebo arms than were seen in the aripiprazole trial. The committee also noted that there appears to be a large placebo effect in the aripiprazole trial, but heard from the manufacturer that the precise cause of this effect is unknown. The committee was aware that, insofar as evidence is available, the CGI and CGAS findings from the trials are not better for aripiprazole than for the comparator treatments.
The committee considered the evidence on adverse events for aripiprazole and each of the comparators presented in the manufacturer's additional analyses. The committee noted that there is substantial variation between the atypical antipsychotics in the adverse events associated with each treatment. The committee was aware of the clinical specialists' view that it is important for adolescents with schizophrenia to have a range of treatment options before considering rescue therapy with clozapine, in order to individualise treatment and to minimise adverse treatment effects. The committee noted that olanzapine is associated with substantial weight gain, as to a lesser extent are quetiapine and risperidone, but that only very small changes in weight gain are seen with aripiprazole. It considered that weight gain may be of considerable importance to adolescents and was concerned that weight gain associated with olanzapine may not be just a short-term problem, but could be a long-term health risk. In terms of changes in prolactin levels, the committee heard from the clinical specialists that risperidone is associated with higher levels of prolactin, as to a lesser extent is olanzapine. Prolactin levels with aripiprazole treatment are generally lower than seen with the other comparator treatments. The committee heard that a change in prolactin level is 1 of a number of contributors to potential sexual dysfunction.
The committee considered the manufacturer's economic model and the critique and exploratory analyses performed by the ERG. It noted that the manufacturer used a decision tree, followed by a Markov model, to estimate the cost effectiveness of first-line aripiprazole compared with first-line olanzapine for the treatment of schizophrenia in adolescents. Data were derived from the manufacturer's adjusted indirect comparison using secondary outcome data for aripiprazole and olanzapine. The committee also considered an updated adjusted indirect comparison from the manufacturer that incorporated risperidone, quetiapine and olanzapine as comparators.
The committee had concerns about a number of aspects of the economic model, including the exclusion of comparators specified in the final scope, primary (PANSS) outcome data and data on relevant adverse events (such as extrapyramidal symptoms and sexual dysfunction). The committee was also aware that the ERG had identified a number of technical errors in the manufacturer's model. The committee heard from the ERG that in the absence of data specific to the population in the scope, data on health state utility at relapse, disutility associated with treatment-related adverse events and resource use assumptions were all derived from studies of adults rather than adolescents.
The committee noted that the manufacturer's initial base-case ICER (provided for the first appraisal committee meeting and following revisions from the ERG) for first-line aripiprazole (in a 3-drug sequence) compared with first-line olanzapine of £6,200 per QALY gained (incremental costs -£69, incremental QALYs 0.004) was based on a number of assumptions that were inappropriate; and that sensitivity analyses conducted by the ERG suggested the ICER could be as high as £233,000 per QALY gained if certain assumptions were varied. Furthermore, it noted that aripiprazole is dominated by risperidone in all of the ERG's exploratory analyses. It concluded that the ICERs presented by the manufacturer could not be accepted without revision. The committee requested further clarification from the manufacturer.
The committee considered the manufacturer's updated economic model that compared sequences of treatments starting with aripiprazole with sequences starting with risperidone. The committee still had concerns about the primary outcome data not being included in the model. It did not agree with the manufacturer's argument that this omission could be justified on the grounds that trial outcomes were not used in ordinary clinical practice; nor did it agree with the manufacturer's reference to NICE's previous guideline on schizophrenia: core interventions in the treatment and management of schizophrenia in adults in primary and secondary care (now replaced by NICE's guideline on psychosis and schizophrenia in adults), in which PANSS scores were used to inform utility values and not considered as a separate outcome measure as an argument for not including PANSS scores in the model. The committee noted that as aripiprazole is associated with smaller changes in PANSS scores than risperidone, olanzapine and quetiapine, the omission clearly favoured aripiprazole. The committee also noted that some adverse events (sexual dysfunction and aggression) were not included in the model, and was aware that there was an error in the manufacturer's adjusted indirect comparison that resulted in the odds ratios of withdrawals (for other reasons) from risperidone compared with withdrawals from aripiprazole being higher in the manufacturer's analysis. The committee noted that the manufacturer's updated base-case analysis shows that treatment sequences in which aripiprazole is used first result in ICERs ranging from £52,750 per QALY gained to £108,800 per QALY gained when compared with treatment sequences in which risperidone is used first. It considered that, in view of the PANSS scores not being included in the model, these ICERs were likely to be underestimated. Furthermore, they are outside the range considered to be a cost-effective use of NHS resources.
In view of the results from the manufacturer's updated base-case analysis and the testimony of the clinical specialists, which highlighted that routine clinical practice is to start treatment with risperidone, the committee concluded that starting treatment with aripiprazole rather than risperidone would not be a cost-effective use of NHS resources.
However, the committee was mindful that in people aged 15 to 17 years with schizophrenia who are intolerant of or have a contraindication to risperidone, or whose schizophrenia has not been adequately controlled with risperidone, the case for aripiprazole is more plausible. The committee considered whether there was any evidence to suggest that aripiprazole should be used ahead of, or only after olanzapine or quetiapine in the treatment pathway for schizophrenia. It noted that the economic analyses suggest little difference between sequences in which aripiprazole precedes olanzapine and vice versa; and although sequences that contain aripiprazole are suggested to be more cost effective than the sequence that contains quetiapine (sequence D), the committee was concerned that the cost of quetiapine was unfairly calculated in the manufacturer's economic model, as optimal packs and doses may not have been considered. The committee agreed that the differences in side effects between these drugs were a more important consideration than the (small) differences in their costs and primary outcomes. Therefore, the committee agreed that aripiprazole should be available on equal terms with other antipsychotic comparators (apart from risperidone), given its good side-effect profile and comparable price to olanzapine and quetiapine
The committee considered whether its recommendations were associated with any potential issues related to equality. The committee was aware that consultees and commentators suggested that 1 area of potential discrimination was that the diagnosis of schizophrenia requires a definitive methodological approach using precise diagnostic criteria detailed in a number of tools, including DSM-IV and K-SADS-PL. The committee noted that although some people with learning difficulties may exhibit psychoses, unless they fulfil the DSM-IV and K-SADS-PL criteria for schizophrenia they do not (by definition) have schizophrenia, and therefore are not appropriate for inclusion in this technology appraisal. It noted that both the DSM-IV and K-SADS-PL criteria are used in clinical practice, as well as in studies of schizophrenia. The committee concluded that there are not sufficient data to provide evidence on how the clinical and cost effectiveness of aripiprazole may differ for people with schizophrenia who have learning difficulties.