Issues in the design of studies for the economic evaluation of new atypical antipsychotics: the ESTO study

Background: Increasing attention is being focused on the costs of healthcare and the need for cost-effective treatments. Drugs for schizophrenia have not escaped this scrutiny, especially now that several new agents are available, with acquisition costs substantially higher than for established therapies. However, most of the existing evaluations of new drugs for schizophrenia have weak designs, either comparing health care costs before and after introduction of the new drug, or being based on modelling approaches incorporating numerous assumptions.

Aim of the Study: The aim of the study was to discuss and resolve the key design issues in the planning of a prospective randomized trial to assess the socio-economic impact of a new atypical antipsychotic (quetiapine).

Methods: Key methodological issues were identified and discussed in the context of the economic evaluation being planned. These were patient recruitment and entry criteria, selection of comparator drug, blinding of doctor and patient, range of socio-economic outcomes, length of follow-up and sample size.

Results: The resulting economic evaluation, the ESTO study, was an international multi-centre randomized controlled trial, with concurrent data collection for a wide range of clinical, economic and quality of life outcomes. The trial had a pragmatic design, enrolling patients experiencing an acute exacerbation on existing therapy. In addition to the presenting exacerbation, patients must have had at least one hospitalization or documented evidence of exacerbation within the previous three years. On admission to the study, existing psychotic medication was withdrawn prior to randomization to quetiapine or haloperidol. Doses of both drugs were titrated up to an optional dose, with flexibility for additional increases if required.

Both patients and doctors were blinded to treatment allocations, on the grounds that, since quetiapine was still in development, unblinded assessments of efficacy would not be credible. Patients were followed for 1 year, irrespective of whether they withdrew from study medication.

A wide range of socio-economic outcomes was assessed, including costs falling on the healthcare sector, other agencies and the family. In addition data were collected on patients’ earnings and quality of life, measured by the Short-Form 36 health profile. Data were also collected on a range of clinical measures, such as the Positive and Negative Syndrome Scale (PANSS), the Clinical Global Impressions (CGI), the AIMS neurological rating scale and the neurological rating scale of Simpson and Angus. This was to assess whether changes in socio-economic end points were indeed matched by changes in the patient’s clinical condition.

Conclusions: The design of studies such as ESTO is inevitably a compromise between control and pragmatism. For example, whilst blinding of doctor and patient may reduce potential bias, this may cause difficulty with compliance owing to the use of additional dummy medications. Despite these compromises, the ESTO study should provide a more reliable assessment of the socio-economic outcomes of a new anti-psychotic and has attracted the widespread support of analysts and investigators. It has already served as a template for other studies and, if the methodology is successful, will have implications for the assessment of similar drugs in the future.