Enhancing therapeutic decisions during robot-assisted gait therapy: Current challenges and development of a novel app-based therapy protocol to address them

Abstract Background Therapists' decisions on influencing patient behavior during robot-assisted gait therapy directly impact clinical practice and, thus, therapeutic efficacy. These manipulations include device parameter adjustments and direct interactions with patients. Little research has been done on how therapists accomplish this task, and knowledge about what constitutes good decisions is scarce. The scarce knowledge is partly due to missing tools to document these decisions. Understanding how therapists assess behavior, manipulate patient behavior, and re-assess behavior for potential improvement could lead to better therapy outcomes. Methods We developed a novel mobile application to document therapeutic decisions based on expert knowledge and literature findings. Each documented manipulation consisted of three parts: The evaluation of the patient behavior (goal-oriented and compensatory), the manipulation by the therapist (direct interaction or device interaction), and the re-evaluation of the behavior concerning the consequences of this manipulation. The resulting application was pilot-tested with 10 patients in 5 sessions and 7 therapists in a real-world clinical setting. Results Therapists selected more than two therapy goals per session for 50% of the training sessions. The results show that 92% of the manipulations were direct interactions and 8% device interactions. Sixteen percent of the direct interactions targeted the selected goal and 43% compensatory movements. The type of manipulation had a statistically significant impact on the rate of improvement at re-evaluation, with direct interactions targeting compensatory movements having the best rate of improvement (48%) and the parameter adjustments having the lowest rate of improvement (25%). Discussion The app developed is a novel approach to documenting therapists' decisions. The data collected in the pilot study revealed several challenges in the decision-making process that therapists can currently only address based on their experience, including goal structure, the relative importance of goal-directed movements versus compensatory patterns, and the type of manipulation taken. Quantitative data provided by the app has the potential to objectify and improve these decisions and could support a transfer of successful treatment approaches between therapists and clinics.