Rural-urban disparities in post-acute therapy utilization: identifying hot spots of low utilization among fee-for-service Medicare beneficiaries toward informing policy and public health interventions

This project has two aims, each one described below.<b><br><br>Aim 1: To quantify rural-urban
disparities in rehabilitation therapy utilization (IRFs/ SNFs/ HHAs) among FFS
Medicare beneficiaries (2021) and evaluate their change over time (2013-2021). </b>Hypothesis 1: Rural Fee For
Service (FFS) Medicare beneficiaries had lower therapy utilization than urban
counterparts in 2021. Hypothesis 2: Rural disparities have been stagnant
or increased as opposed to significantly reduced over time (2013-2021),
stratified for the pre- (2013-2019) and post-pandemic (2020-2021) times. We will use hierarchical linear multiple regressions. Dependent
variable:<b> </b>counties’<b> therapy utilization rate for </b>IRFs,
SNFs, and HHAs combined. Independent variable<b>:
rural area</b>, per two indicators: a) rural residency of FFS beneficiaries b)
rural county gradient. Covariates: 1) FFS beneficiaries’
characteristics; 2) FFS Medicare expenditures; 3) counties’ disability
statistics, e.g., poverty; 4) community-level health, health access and social
determinants of health; and 6) regions and states. The significance of the
interaction between years and rurality will be tested. <br><b></b>

<b><br>Aim 2: To build interactive,
user-centered maps of rehabilitation therapy utilization (IRFs / SNFs / HHAs), identifying
hot spots of low utilization (in 2021) and their evolving trends
(2013-2021).</b> A GIS (ArcGIS Pro) will be used to
spatiotemporally analyze the data used in the Aim 1. First, we will develop <b>choropleth
maps</b>: gradients of county-level utilization rates, intersected with
rural areas. Second, <b>hot spot analyses</b> (statistical spatial
clustering) will map clusters of counties with low utilization. Third, a <b>spatiotemporal</b>,
<b>emerging hot spot analysis</b> (2013-2021) will map<b> </b>areas
with up to eight types of time-trends such as those showing an<b> intensifying
or persistent low utilization</b>. The spatiotemporal analysis will be
also stratified
for the pre- (2013-2019) and post-pandemic (2020-2021) time periods. All the maps, with customizable options, will be shared online for
public access. An<b> Advisory Group</b> of target end-users (e.g., disability
advocates, public health agents) will provide input throughout to design the maps’
attributes and refine them after beta testing