Sample size calculation of clinical trials in geriatric medicine
- PDF / 547,226 Bytes
- 4 Pages / 595.276 x 790.866 pts Page_size
- 85 Downloads / 214 Views
REVIEW
Sample size calculation of clinical trials in geriatric medicine Graziella D’Arrigo1 · Stefanos Roumeliotis1,2 · Claudia Torino1 · Giovanni Tripepi1 Received: 27 February 2020 / Accepted: 11 May 2020 © Springer Nature Switzerland AG 2020
Abstract A preliminary step when planning a randomized clinical trial (RCT) is the sample size calculation. This is the determination of the optimal number of patients which ensures an adequate power to the study to detect as statistically significant a certain between-arms difference, if any, in the frequency/magnitude of a specific endpoint. The sample size calculation is performed by specific calculators requiring as input variables the expected effect size, the alpha error (α), the beta error (β) and the allocation ratio, this latter being the ratio between the number of participants allocated to the arms of a RCT. Herein, we provide a series of examples of sample size calculation in the context of superiority RCTs in elderly. Keywords Randomized clinical trial · RCT · Sample size calculation · Alpha error · Beta error · Study power
Introduction A randomized controlled clinical trial (RCT) is an experimental study used by researchers to compare a given outcome between two or more groups of patients assigned to different treatments and followed up over time [1]. The determination of the optimal sample size is a fundamental, preliminary step of an RCT, because it ensures an adequate power to detect as statistically significant a certain betweenarms difference, if any, in the frequency/magnitude of a specific endpoint. The sample size calculation also addresses some practical issues. In fact, the enrolment of too many patients exposes more participants than needed to the intrinsic risks of the experiment (for example, adverse events), whereas if the RCT is not adequately powered (i.e., it enrolls a non-sufficient number of participants) it will be considered inconclusive, particularly in case of results which did not attain the statistical significance. RCTs can be used to assess superiority, equivalence, and non-inferiority of an experimental drug versus a comparator. Herein, we provide * Giovanni Tripepi [email protected] 1
Institute of Clinical Physiology (IFC‑CNR), Clinical Epidemiology and Physiopathology of Renal Diseases and Hypertension of Reggio Calabria, Ospedali Riuniti, via Vallone Petrara, Reggio Calabria, Italy
Division of Nephrology and Hypertension, 1st Department of Internal Medicine, School of Medicine, AHEPA Hospital, Aristotle University of Thessaloniki, Thessaloníki, Greece
2
a series of examples of sample size calculation in the context of superiority RCTs performed in the elderly and by considering binary outcomes. The calculations of sample size for non-inferiority and equivalence trials are reported elsewhere [2].
Type 1 and type 2 errors To calculate the sample size, two main hypotheses should be preliminary fixed. In a superiority RCT [3] comparing two hypothetical drugs (A versus B), the null hypothesis (H0) is that drug A is
Data Loading...
