Construct Validity Supported for Multiple DMOs

A recent webinar hosted by members of the Mobilise-D consortium revealed that construct validity was supported for multiple digital mobility outcomes (DMOs). The clinical validation studies involved almost 2,400 patients and evaluated 24 DMOs across 4 participant cohorts: chronic obstructive pulmonary disease (COPD), multiple sclerosis (MS), Parkinson’s (PD), and proximal femur fracture (PFF).

Construct validity testing assessed convergent, divergent, and known-groups validity against a primary hypothesis set for each DMO and construct*. Results for the 4 conditions assessed as part of the Mobilise-D clinical validation studies were as follows.

• COPD: construct validity supported for 17 / 24 DMOs

• MS: construct validity supported for 19 / 24 DMOs

• PD: construct validity supported for 13 / 24 DMOs

• PFF: construct validity supported for 17 / 24 DMOs

What is particularly encouraging is that construct validity was supported for 10 of the DMOs across all four conditions. The suggests that the DMOs could have explanatory power across many other mobility impacting conditions. The table below sets out the results across all DMOs and conditions.

Insights from a Large-Scale Survey of Study Participants

Enoda founder Dr. Alison Keogh also provided insights from a large survey of study participants with almost 1,500 responses. Participants were asked about the comfort of the sensor used to capture the data, as well as the aspects of the protocol. The survey** found that:

• 92% of participants agreed or strongly agreed that remote monitoring was acceptable to them;

• 85% agreed or strongly agreed that the way the device was fastened was acceptable to them; and

• the results were consistent across the cohorts, and there was no evidence that factors such as activity levels, disability or education had any impact on the scores.

Dr. Keogh also provided insights from work with the participant group from Mobilise-D on the importance of mobility and their acceptance of mobility monitoring. She noted that:

• walking is a meaningful aspect of health that patients care about, and is integral to their sense of self;

• participants felt that mobility is not given enough importance from healthcare professionals;

• if patients understand why they are wearing a device and can see value in it, they will support it; and

• working with patients to provide custom visualisations is an important aspect promoting understanding of the data and gaining acceptance of mobility monitoring.

Conclusion

The latest Mobilise-D findings represent a major milestone for real-world mobility assessment. Demonstrating that construct validity was supported for a substantial number of DMOs across COPD, MS, PD and PFF provides strong evidence that these measures capture meaningful aspects of mobility performance across diverse clinical populations.

Equally encouraging is the feedback from participants. High levels of acceptability, comfort, and engagement underline that real-world mobility monitoring is not only technically feasible, but is also accepted and indeed welcomed by patients. They care deeply about their mobility, and want to see it given greater prominence in their health care. The Mobilise-D DMOs are increasingly well positioned to become robust, patient-centred tools for understanding disease progression, informing clinical decisions, and shaping the future of mobility-focused healthcare.

Notes

*Construct validity testing.

Convergent validity tests assessed the relationships of the DMOs to constructs that were hypothesised to be related (for example that the DMOs were correlated with clinical scales such as the EDSS amongst MS participants or the UPDRS for PD).

Divergent validity tests examined relationships against constructs that are hypothesised to not be related (e.g. correlations with tremor severity for PD participants).

Known-groups validity tests compared DMO values across categories of relevant constructs for which they are expected to differ (for example it was hypothesised that DMO values for PD participants in Hoehn & Yahr stage 1 = 2, but 1 ≠ 3 and 2 ≠ 3).

**The survey questionnaire was developed with study participants and used a 5-point Linkert scale ranging from ‘Strongly Disagree’ to ‘Strongly Agree’.