Consensus on the clinical utility of digital mobility outcomes for personalized clinical decision support in parkinson's disease.

Background: Digital mobility outcomes (DMOs) have emerged as novel biomarkers offering objective, quantitative, and examiner-independent outcome measures for clinical studies. Unfortunately, research efforts on DMOs have not yet investigated the domain of clinical utility in Parkinson's disease, i.e. providing evidence of improvements in health outcomes, diagnosis, decision-making, or prevention when compared to e.g. standard-of-care procedures. This manuscript, via a consensus building approach, aims to create a structured conceptual framework to map the knowledge generated by DMOs with clinical domains that could benefit from it.

Methods: We conducted a three-round consensus-building study with 12 experts recruited from the Mobilise-D consortium's Parkinson's Disease Working Group. The experts designed and ranked different aspects of the conceptual framework via a 5-level Likert scale for level of agreement. Consensus for the different points evaluated was based on a double threshold: the simultaneous presence of a high level of agreement had to be accompanied by a low level of disagreement. As secondary objectives, the experts were asked to rate the practical application of DMOs by evaluating the timeline to applicability, the foreseen challenges for their implementation in clinical settings, and their main role in the decision-making process.

Results: A full consensus on the clinical utility framework was achieved after three rounds. The final framework consisted of three main categories (Disease Diagnosis, Patient Evaluation, and Treatment Evaluation) and six underlying domains (Enhancing Diagnostic Procedure, Predicting Risk, Timely Detecting Deterioration, Enhancing Clinical Judgment, Selecting Treatment, and Monitoring Treatment Response). The experts believed in the next 1-5 years DMOs will play a relevant role in clinical decision making, complementing care knowledge with useful digital biomarkers information. However, the main challenge to address is the definition of clear reference value for DMOs interpretability.

Conclusions: This framework provides a structure for subsequent studies to build into by diversifying expert cohorts and expand our findings beyond PD. Additionally, our results support researchers planning future clinical trials where DMOs can play a valuable role for clinical decision support. Ultimately, this is the first step toward developing guidelines to assess DMOs' clinical utility and support their integration into Real World clinical practice.