Making the most of a few small population clinical trials

"Patients in rare diseases should be entitled to the same quality of treatment as other patients". The recent implementation of international and national strategy plans for rare diseases has led to an incredible increase of data quantity. As is generally known, such data become available through various sources, for instance; claim databases, e-health records, national registries and experimental studies. Information provided from either of these sources can be utilized during the evaluation of a (new) treatment.

Depending on the stakeholder, we often need to evaluate multiple questions; such as, (i) do the treatment's risks outweigh the added benefit?, (ii) does the treatment's added benefit outweigh the costs?, (iii) does the treatment help each and every individual patient or (iv) does the treatment actually work? The latter question is often evaluated by a randomized controlled trial. According to Orphanet, more than 1829 trials among 29 countries focus on more than 800 rare conditions. This number of historical and ongoing trials explains the urgent need for evaluating and developing tailor-made statistical methods for small populations and so far, three recent large European projects responded to materializing this need (ASTERIX,IDEAL,INSPIRE).

Orphan drugs for rare diseases are usually investigated through multinational randomized controlled trials. Investigations in such a global setting may lead to increased inconsistency, given that the clinical expertise, the standards of care and the used facilities vary e.g. between each country. Non-homogeneous data may make single trials less reliable, while smaller trials are often more feasible to conduct. This means that the need to explore methods for the synthesis of a few small studies through meta-analysis is even more necessary. This thesis focuses on the latter.