It is with great interest that we read the article "Assessing Effectiveness and Costs in Robot-Mediated Lower Limbs Rehabilitation: A Meta-Analysis and State of the Art" by Carpino et al. , which was published in the Journal of Healthcare Engineering, Volume 2018 (ID 7492024). We believe the authors are investigating and discussing a very important topic, i.e., the financial efficiency of modern technology in rehabilitation.
We, however, would like to express our concern about some of their calculations and conclusions.
1. Purchasing Costs
Our biggest concern is the indicated purchasing costs. In their calculations, Carpino et al.  differentiate between operational machines (commonly referred to as end-effector devices) and wearable robots (commonly referred to as exoskeleton devices) with the goal to compare the financial efficiency of the two classes with each other and with conventional training. The authors have based all their calculations on a purchasing price of 330,000 [euro] for a Lokomat (as an example of an exoskeleton) and 30,000 [euro] for a Gait Trainer GT1 (as an example of an end-effector). However, the LokomatPro is a state-of-the-art representative of the exoskeleton group (and at the mentioned price in Italy, it likely includes the optional FreeD feature as well as more than one year of maintenance). The GT1, which they use as the representative of the end-effector device group, has long been replaced by the GTII and further developed into the G-EO, which has a list price of 250,000 [euro]. Therefore, when comparing a state-of-the-art exoskeleton device to a state-of-the-art end-effector device, the difference in costs is not, as indicated, tenfold, but much smaller. Similarly, the maintenance costs, which in the article are indicated as 10% of the device or 22,500 [euro] per year, are actually just below 10,000 [euro] per year for the Lokomat based on the data provided by Hocoma device manufacturer and just slightly higher than that quoted by the G-EO manufacturer. We cordially invite the authors to repeat their calculations with these numbers.
2. Human Resources: Establishing Efficient Settings
Secondly, we have concerns about the necessary human resources stated by the authors. They indicate that 1.19 therapists are necessary to conduct a conventional therapy session, while 1 therapist is used to conduct a robotic session, regardless of the type of the robotic device. This stands in contrast with previously published data by Morrison  and Esquenazi et al. . Morrison  indicated that 1 trainer is required for robotic locomotor training, while up to four trainers are needed for a manual locomotor training session. This difference is likely due to the fact that Morrison compared intensive, repetitive locomotor training assisted by a robot (robotic locomotor training, rLT) to similarly intensive, repetitive locomotor training assisted manually by physical therapists and physical therapy aids (manual locomotor training, mLT). Esquenazi et al....