A programme evaluation of 'First Steps': A peer-conceived, developed and led self-management intervention for people after a Parkinson's diagnosis.

The study did not recruit to target: it aimed for 80 participants but enrolled 57, so the between-group comparison was weaker than planned. The design changed from an intended randomised trial to a pragmatic step wedge, and allocation was by location rather than random assignment. Groups were not comparable in location or deprivation: control participants all lived in urban locations, many in major conurbations, and lived in less deprived areas. Baseline differences were present: the intervention group reported more problems with usual activities on EQ5D (p = 0.035), and anxiety, depression and physical activity scores trended higher in the intervention group. Post-intervention timing was inconsistent because 11 people had not completed both days by 12 weeks, so 24-week data were used for them. Only 22 of 36 people allocated to the intervention completed both days, and 8 did not start without giving a reason. Fidelity was not assessed at Hampshire, and at Devon a mental health discussion was omitted and a slide removed. Participants were largely active, affluent and had low anxiety or depression, so results may not apply to people with more severe symptoms or from more deprived areas. The evaluation took place before the pandemic and First Steps has since moved online, so these face-to-face results may not translate directly.

The evaluation was funded by Parkinson's UK, which also intended national adoption of the programme. The authors declared no potential conflicts of interest. Some authors were supported by NIHR centres.

Do not read the within-group increase in physical activity as proof that First Steps works better than usual care. The between-group comparison found no significant differences, and the control group tended to be more active at baseline.