Show simple item record

dc.contributor.authorRowe, Gina C.
dc.date.accessioned2014-01-22T15:04:16Z
dc.date.available2014-07-09T12:07:56Z
dc.date.issued2013
dc.identifier.urihttp://hdl.handle.net/10713/3652
dc.descriptionUniversity of Maryland, Baltimore. Nursing. Ph.D. 2013en_US
dc.description.abstractBackground: Potentially preventable hospital admissions (PPAs) and emergency department (ED) visits (PPVs) are those that might have been prevented if patients had received better primary care. A significant number of ED visits in the United States and about a third of those in Maryland are "ambulatory care sensitive," or potentially preventable. Geographic variation in PPV rates reflects community-level differences in primary care access, social determinants of health-seeking behavior, and health disparities. Higher rates are noted in poor communities and vulnerable populations. Purpose: To compare and explain the geographic variance in Maryland PPV rates for total and uninsured populations and test the predictive value of regression models developed using generalized linear regression and geographic information systems. Analysis of geographic variance in PPV rates across the Baltimore metropolitan statistical area (MSA) used neighborhood-level social determinants to determine whether social capital can mediate the negative impact of living in a disadvantaged neighborhood on PPV rates. Methods: Two cross-sectional, ecologic regression analyses of secondary data aggregated to the zip code tabulation level were conducted. Generalized linear and geographic regression models were built using SPSS and ArcGIS statistical software, and results were compared to determine which model(s) best explained geographic variance in PPV rates. Social capital measures were obtained from the Baltimore Ecosystem Study. Results: In Maryland, geographic hot spots of increased PPV rates were highly correlated for uninsured and total populations, but uninsured PPV rates were more clustered in urban areas. Poisson and geographically weighted regression (GWR) models explained the most PPV rate variance. Significant predictors were per capita income, female-headed households, and level of education. In the Baltimore MSA, Poisson and GWR models predicted 85-86% of PPV rate variance; relative poverty and female-headed households were significant predictors but percent uninsured and per capita primary care physicians were not. Social capital was a significant partial mediator of all measures of neighborhood disadvantage reviewed. Conclusion: Communities with high social capital may offer health-protective benefits to residents, even mediating the negative impact of living in a disadvantaged neighborhood. Reducing PPVs requires consideration of population-level health-seeking behaviors and promotion of neighborhood-level social capital, particularly for single mothers.en_US
dc.language.isoen_USen_US
dc.subjectambulatory care sensitive conditionsen_US
dc.subjectgeographic varianceen_US
dc.subjectneighborhood disadvantageen_US
dc.subjectsocial capitalen_US
dc.subject.lcshNeighborhoods--health aspectsen_US
dc.subject.meshEmergency Medical Servicesen_US
dc.subject.meshHealth Services Accessibilityen_US
dc.subject.meshMarylanden_US
dc.subject.meshPrimary Health Care--utilizationen_US
dc.subject.meshSocial Determinants of Healthen_US
dc.titleQuantifying Neighborhood-Level Social Determinants of Potentially Preventable Emergency Department Visits in Marylanden_US
dc.typedissertationen_US
dc.contributor.advisorJohantgen, Mary E.
dc.identifier.ispublishedNoen_US
dc.description.urinameFull Texten_US
refterms.dateFOA2019-02-19T18:05:13Z


Files in this item

Thumbnail
Name:
Rowe_umaryland_0373D_10492.pdf
Size:
2.998Mb
Format:
PDF

This item appears in the following Collection(s)

Show simple item record