EVITA: a tool for the early evaluation of pharmaceutical innovations with regard to therapeutic advantage.

BACKGROUND: New drugs are generally claimed to represent a therapeutic innovation. However, scientific evidence of a substantial clinical advantage is often lacking. This may be the result of using inadequate control groups or surrogate outcomes only in the clinical trials. In view of this, EVITA was developed as a user-friendly transparent tool for the early evaluation of the additional therapeutic value of a new drug. METHODS: EVITA does not evaluate a new compound per se but in an approved indication in comparison with existing therapeutic strategies. Placebo as a comparator is accepted only in the absence of an established therapy or if employed in an add-on strategy on top. The evaluation attributes rating points to the drug in question, taking into consideration both therapeutic benefit and risk profile. The compound scores positive points for superiority in efficiency and/or adverse effects as demonstrated in randomized controlled trials (RCTs), whilst negative points are awarded for inferiority and/or an unfavorable risk profile. The evaluation follows an algorithm considering the clinical relevance of the outcomes, the strength of the therapeutic effect and the number of RCTs performed. Categories for therapeutic aim and disease severity, although essential parts of the EVITA assessment, are attributed but do not influence the EVITA score which is presented as a color-coded bar graph. In case the available data were unsuitable for an EVITA calculation, a traffic-type yield sign is assigned instead to criticize such practice. The results are presented online http://www.evita-report.de together with all RCTs considered as well as the reasons for excluding a given RCT from the evaluation. This allows for immediate revision in response to justified criticism and simplifies the inclusion of new data. RESULTS: As examples, four compounds which received approval within the last years were evaluated for one of their clinical indications: lenalidomide, pioglitazone, bupropion and zoledronic acid. Only the first achieved an EVITA score above zero indicating therapeutic advantage. CONCLUSIONS: The strength of EVITA appears to lie in its speedy assessment of the potential therapeutic advantage of a new drug for a given indication. At the same time, this approach draws attention to the typical deficits of data used for drug approval. EVITA is not intended to replace classical health technology assessment reports but rather serves as a screening tool in the sense of horizon scanning.

Hospitalization rates of generic metoprolol compared with the original beta-blocker in an epidemiological database study.

PURPOSE: A less favourable galenic profile of generic formulations of the beta-blocker metoprolol raised the concern of a higher risk for serious cardiovascular (CV) events. We assessed hospital admission rates for CV diseases and prescription prevalences of various drugs using claims data of statutory health insurances (SHIs) to compare the incidence of serious CV events among users of original and generic metoprolol. Index events included hospitalization due to myocardial infarction, hypertensive crisis and stroke. METHODS: Data files of three SHIs were linked with dispensing data of drug prescriptions from each pharmacy's electronic data processing centre on an individual basis. Incidences of hospital admissions among patients receiving original metoprolol and among patients treated with the generic equivalent were compared by logistic regression, stratified for Bremen and the rest of Northern Germany. Risk estimates and confidence intervals were adjusted for confounders. RESULTS: A total of 49,673 patients receiving metoprolol were identified within a cohort of 3,649,285 insurance members. While the crude analysis revealed a higher risk for index events in patients receiving the generic drug (Bremen: RR 1.45; Northern Germany: RR 1.14), no elevated risk remained after confounder adjustment (Bremen: OR 1.06; Northern Germany: OR 1.04). Among co-morbid conditions considered as confounders, a previous CV event and an elevated thromboembolic risk exerted the strongest effect on index events. CONCLUSIONS: SHI data are a valuable source for pharmacoepidemiology and health services research in Germany. Incidence rates of serious CV events did not reveal any noticeable differences between the original and the generic group after confounder adjustment.