TY - JOUR
T1 - Spatial Non-stationarity in Opioid Prescribing Rates
T2 - Evidence from Older Medicare Part D Beneficiaries
AU - Kim, Seulki
AU - Shoff, Carla
AU - Yang, Tse Chuan
N1 - Funding Information:
The views expressed in this article are those of the authors, and no official endorsement by the Department of Health and Human Services or the Centers for Medicare & Medicaid Services is intended or should be inferred.
Publisher Copyright:
© 2019, Springer Nature B.V.
PY - 2021/4
Y1 - 2021/4
N2 - Previous research that examined spatial patterns of opioid prescribing rates and factors associated with them has mainly relied on a global modeling perspective, overlooking the potential spatial non-stationarity embedded in these associations. In this study, we investigate whether there are spatially non-stationary associations between opioid prescribing rates and key characteristics of older Medicare Part D beneficiaries and their prescribers using several data sources from the Centers for Medicare and Medicaid Services. All measures are aggregated to the ZIP code-level and a total sample size of 18,126 ZIP codes is included in the analyses. Our descriptive results from geographically weighted regression and the Monte Carlo significance test suggest that most of the associations between the characteristics of beneficiaries and prescribers and opioid prescribing rates are spatially non-stationary. Our findings not only challenge the conventional analytic approach by highlighting the importance of a local modeling perspective in opioid prescribing research, but also offer nuanced insight into how opioid prescribing rates are related to possible determinants across space.
AB - Previous research that examined spatial patterns of opioid prescribing rates and factors associated with them has mainly relied on a global modeling perspective, overlooking the potential spatial non-stationarity embedded in these associations. In this study, we investigate whether there are spatially non-stationary associations between opioid prescribing rates and key characteristics of older Medicare Part D beneficiaries and their prescribers using several data sources from the Centers for Medicare and Medicaid Services. All measures are aggregated to the ZIP code-level and a total sample size of 18,126 ZIP codes is included in the analyses. Our descriptive results from geographically weighted regression and the Monte Carlo significance test suggest that most of the associations between the characteristics of beneficiaries and prescribers and opioid prescribing rates are spatially non-stationary. Our findings not only challenge the conventional analytic approach by highlighting the importance of a local modeling perspective in opioid prescribing research, but also offer nuanced insight into how opioid prescribing rates are related to possible determinants across space.
KW - Geographically weighted regression
KW - Medicare Part D prescription drug event
KW - Opioid prescribing rate
KW - Spatial non-stationarity
UR - http://www.scopus.com/inward/record.url?scp=85076572055&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85076572055&partnerID=8YFLogxK
U2 - 10.1007/s11113-019-09566-7
DO - 10.1007/s11113-019-09566-7
M3 - Article
AN - SCOPUS:85076572055
SN - 0167-5923
VL - 40
SP - 127
EP - 136
JO - Population Research and Policy Review
JF - Population Research and Policy Review
IS - 2
ER -