Figure 4

Fictitious survival analyses explaining why a comparison of two different treatments without an untreated control group does not make sense: When one group of patients undergoing the ‚nuts' treatment does not benefit from this, but is compared to the ‚plums' treatment group increasing the curvature angle, there surely may be differences in controlled studies (randomized or not). But it does not show one of the interventions is really beneficial to the patient ('orange'). So, a controlled study design without an untreated control group is not providing any evidence for an intervention as investigated with the help of this study design [as demonstrated in 16]. With kind permission of Pflaum, Munich (Weiss HR: Best practice in conservative scoliosis care. 4th edition in press).