Community socioeconomic status and public access defibrillators: A multilevel analysis

Background: Although current guidelines recommend that distribution of public-access defibrillators (PADs) should take into account area-level risk of out-of-hospital cardiac arrest (OHCA), community socioeconomic status (SES) can unduly influence policy implementation in positioning PADs. Using recent, complete data from Seoul Metropolitan City, Korea, this study aims to examine whether community SES is associated with distribution of PADs, in terms of per capita count and risk-grid coverage. Methods: A cross-sectional, observational study was conducted using three sources of administrative data: (1) PAD registry data (2007-2015), (2) OHCA database (2010-2014), and (3) community socioeconomic characteristics of two sub-city levels (neighborhoods nested in districts). We examined the relationship between neighborhood per capita tax, an SES proxy, with each of the two outcome variables. After examining per capita number of PADs and risk-grid coverage by neighborhood tax quartile, multilevel linear regression analysis was conducted to account for the nested nature of data and also to control for OHCA risk in the model. Results: A total of 6609 PADs in 405 neighborhoods were included in the analysis. The average number of positioned PADs per 10,000 persons was 7.45, showing a gradient by neighborhood SES (4.92 in the lowest SES quartile vs 12.66 in the highest). Risk-grid coverage was around 10% across all neighborhood SES quartiles. These findings remained valid in the multilevel analysis: per capita number of PADs was still positively associated with neighborhood SES, while risk-grid coverage of PADs was not. Conclusions: More affluent neighborhoods in Seoul exhibit higher per capita PADs, even accounting for OHCA risk, while risk-grid coverage is generally low regardless of community SES. Seoul's ongoing program aimed to increase PAD coverage should also pay attention to improving community-level inequality as well as distributional efficiency. (C) 2017 Elsevier B.V. All rights reserved.