Background Power of attorney (POA) is a legal document in which one individual gives authority to another to act or make welfare and health care decisions for them, in the event that they become incapacitated. A public awareness campaign, first implemented in Glasgow City, and rolled out to other parts of Scotland in stages, was previously found to increase POA registrations. The current study measures the impact of the campaign on registrations and applications to legal assistance, as a proxy for low socioeconomic (SES) status applications.

Methods POA registration and legal aid application (LAA) for POA data were analysed between January 2010 and December 2018. Multilevel Negative Binomial models for POA and LAA registrations nested by council and annual quarter were run using RIGLS estimation in MLwin, adjusting for pre and post intervention period for each locality, campaign (variable ranging between 0–3 dependent on intensity of campaign measured by the number of media platforms received), and offset term mid-year population estimate for those aged 65 year+. A further model was then run for outcome LAA registrations with offset term total registrations, in a similar way, to examine the impact of the campaign on the proportion of registrations with legal assistance.

Results In Glasgow City POA registrations rose by 33.3% between 2013 and 2014, following the introduction of the campaign, compared with 17.3% in the rest of Scotland. LAA during this period rose by only 10.6% in Glasgow and 16.3% in the rest of Scotland. However, when the data for the whole study period were modelled, the relative risk of a registration for those living in an area with the full campaign was RR=1.12 (1.07, 1.17) those living where no campaign was in place. Relative Risk rose in an approximate stepwise fashion with increasing campaign intensity. Relative risk of a LAA for the same group was 1.10 (1.01, 1.21). When LAA proportion of registrations was instead modelled over time, ie LAA as outcome with total registrations as an offset, the campaign variable was not significantly associated with the outcome.

Conclusion During the period of the campaign, area-level increases in LAA were associated with the timing, intensity and location of the media campaign, in a similar way to that of all POA registrations. This suggests that the campaign increased registrations from both lower socioeconomic backgrounds and all other SES in a similar way, therefore neither reducing nor increasing inequalities in POA registrations.