|| Checking for direct PDF access through Ovid
There has been considerable debate in recent years about whether, and how, to risk-adjust quality measures for sociodemographic characteristics. However, geographic location, especially rurality, has been largely absent from the discussion.To examine differences by rurality in quality outcomes, and the impact of adjustment for individual and community-level sociodemographic characteristics on quality outcomes.The 2012 Medicare Current Beneficiary Survey, Access to Care module, combined with the 2012 County Health Rankings. All data used were publicly available, secondary data. We merged the 2012 Medicare Current Beneficiary Survey data with the 2012 County Health Rankings data using county of residence.We compared 6 unadjusted quality of care measures for Medicare beneficiaries (satisfaction with care, blood pressure checked, cholesterol checked, flu shot receipt, change in health status, and all-cause annual readmission) by rurality (rural noncore, micropolitan, and metropolitan). We then ran nested multivariable logistic regression models to assess the impact of adjusting for community and individual-level sociodemographic characteristics to determine whether these mediate the rurality difference in quality of care.The relationship between rurality and change in health status was mediated by the inclusion of community-level characteristics; however, adjusting for community and individual-level characteristics caused differences by rurality to emerge in 2 of the measures: blood pressure checked and cholesterol checked. For all quality scores, model fit improved after adding community and individual characteristics.Quality is multifaceted and is impacted by individual and community-level socio-demographic characteristics, as well as by geographic location. Current debates about risk-adjustment procedures should take rurality into account.