Health Policy & Management Theses and Dissertations

Permanent URI for this collection


Recent Submissions

Now showing 1 - 5 of 43
  • Item
    (2023) Eschenlauer, Adam; Franzini, Luisa; Health Services Administration; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    Background: Polypharmacy is a growing issue that affects individuals of all ages yet is most prevalent among patients aged 65 and older with chronic comorbidities. Although integral to most treatment plans, pharmaceutical intervention may negatively impact one’s health when five or more medications are taken daily. Given the concurrent rise in elderly population and polypharmacy prevalence, it is vital that we better understand the impact that concomitant medication use has on this vulnerable segment of population.Purpose: This research examines the factors leading to polypharmacy among the elderly population and explores its various impacts on healthcare utilization. Data and Methods: This study uses Medical Expenditure Panel Survey (MEPS) Data. Fixed-Effects regression analyses examine relationships between predictive factors and polypharmacy, polypharmacy and expenditures, and polypharmacy and utilization. Classification models assess the ability of machine learning to correctly predict utilization within the sample population. Key Results: Aside from clinical indicators, demographic and socio-economic factors play a role in determining polypharmacy status. Polypharmacy risk is higher for women (1.088, p < 0.001), high income individuals (1.107, p < 0.01), and those covered by Medicaid (1.110, p < 0.001). Conversely, married individuals (0.930, p < 0.001) and non-Hispanic Blacks (0.864, p < 0.001) have reduced risks of polypharmacy. We find polypharmacy to be associated with higher total (p < 0.001), inpatient (p < 0.01), outpatient (p < 0.01), and prescription medical expenditures (p < 0.001) when holding other predictors constant. We find the risk of hospitalization to be higher for polypharmacy patients (RR: 1.592, p < 0.001) than nonpolypharmacy patients after controlling for multimorbidity and medication class. Lastly, machine learning algorithms classify admissions with an overall accuracy of 84.9%; however, a low true positive rate (TPR) of 41.7% and high true negative rate (TNR) of 96.5% indicate best performance is achieved in predicting non-admissions. Conclusion: Polypharmacy is associated with several non-clinical factors and has a statistically significant impact on medical expenditures and admissions. Though imperfect, predictive analysis methods improve our ability to identify patients at risk for admissions and present a potential opportunity for future applications aimed at reducing utilization and costs.
  • Item
    Home Hemodialysis Utilization and Health Outcomes among Racial and Ethnic Minority Populations
    (2023) Zhu, Ying; Franzini, Luisa; Health Services Administration; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    Background: Home hemodialysis (HHD) offers end-stage renal disease (ESRD) patients greater flexibility and advantages in terms of health outcomes over in-center hemodialysis. There is limited research on the differences between home and center hemodialysis (CHD) and preferences among racial/ethnic minorities, despite ESRD disproportionately affecting these groups. Methods: This project aims to explore the usage and health outcomes of HHD vs. CHD with a focus on racial/ethnic differences using a systematic review of the global main academic database from 2004 to 2022 (study 1), logistic regression and negative binomial analysis of the U.S. Renal Disease System (USRDS) cumulative core data since 2010 merged with 2016-2019 Medicare clinical claims (study 2), and qualitative research using semi-structured interviews with 18 nephrologists and 5 other hemodialysis providers in 8 states of the U.S. (study 3). Results: Study 1: from 3,114 unique studies, six studies met the inclusion criteria and all of them were comparative cohort studies; five out of six studies with a total of 3,172 White patients (68%) and 1,477 minority patients (32%) reported the utilization of HHD; in four of the six studies, the adjusted odds ratio for HHD treatment was shown to be significantly lower for patients of racial or ethnic minorities than for White patients; three out of six studies examined racial/ethnic differences in mortality and other outcomes indicating a lower risk of death for minorities in home hemodialysis. Study 2: minorities were significantly less likely to use HHD than Whites; most minority patients were younger and had fewer comorbidities than Whites, and all minority groups displayed significantly lower mortality and hospitalization incidences than the White group with adjustment on multiple covariates; in the overall and main racial/ethnic cohorts, HHD showed a significantly lower risk of death than CHD after confounding for major risk factors. Study 3: the majority of the interviewees felt that HHD was a viable, safe, and most cost-effective treatment for those with kidney failure, it offered many advantages over traditional CHD but there is a need for additional training and support for the patient, family, provider; minorities and White patients differed in their attitude toward dialysis care, social norms on HHD, and perceived control of personal health. Conclusion: There were major obstacles and considerable racial/ethnic variations in HHD utilization and health outcomes in the US. This study showed that the promotion of HHD will probably require a systematic overhaul in kidney disease management and education.
  • Item
    (2023) Omeaku, Nina; Boudreaux, Michel; Health Services Administration; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    This dissertation examines the impact of cancer on financial outcomes for individuals aged 50 and older and documents how this relationship varies by sex and race/ethnicity. I then turn to the impact of depression on out-of-pocket medical spending among those with a history of cancer. Findings suggest cancer can have a deleterious effect on the financial outcomes of those who are diagnosed with cancer. Out-of-pocket spending rises in the year of diagnosis, reduced earnings persist beyond diagnosis, and depression increases out-of-pocket spending. I fail to find evidence that the relationship between cancer and financial outcomes is moderated by sex or race-ethnicity.
  • Item
    (2023) Wang, Nianyang; Chen, Jie; Health Services Administration; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    There are currently over 6.2 million people with Alzheimer’s Disease and Related Dementias (ADRD) in the US and this population will grow to over 13 million by 2060. Patients with ADRD have high rates of health care utilization due to their complex health needs, which may have been impacted by the COVID-19 pandemic.This dissertation examined the patterns of health care utilization during COVID-19 for the ADRD population. I used 2019-2020 claims data from Inovalon MORE2 to identify emergency department (ED) visits, preventable ED visits, hospitalizations, potentially preventable hospitalizations (PPH), office/clinic visits, and telehealth visits and the Fall 2020 and Winter 2021 Medicare Current Beneficiary Survey COVID-19 Supplements to study telehealth access before and during COVID-19 and telehealth use during COVID-19. The first aim examined patterns in health care utilization before and during COVID-19 using an interrupted time series analysis. The results showed that the total number of ED visits decreased while the percent of preventable ED visits stayed the same, the total number of hospitalizations and the percent of PPH decreased, and office/clinic visits decreased while the percent of telehealth visits increased during COVID-19 for patients with ADRD. The second aim examined the impact of COVID-19 on the rate of monthly PPH for Managed Medicaid enrollees with ADRD and Medicaid expansion status using a difference-in-differences design. The results showed that COVID-19 did not affect the rate of PPH for patients with ADRD by state Medicaid expansion status. The third aim evaluated telehealth access and use for Medicare enrollees with a focus on enrollees with ADRD and Medicare Advantage within the context of COVID-19. Medicare enrollees with ADRD did not have different rates of telehealth access before or during COVID-19 or telehealth use during COVID-19 compared to their non-ADRD counterparts. The interaction of Medicare Advantage and ADRD was not significant in rates of telehealth access or use. Findings from this dissertation provide policymakers with evidence on how to predict and prepare for the health care needs of future pandemics for vulnerable populations such as patients with ADRD.
  • Item
    (2023) Buchongo, Portia; Franzini, Luisa; Health Services Administration; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    Despite substantial federal and state investments made to promote the National Diabetes Prevention Program (NDPP) as a population health strategy for diabetes prevention it remains underutilized. Research has pointed to a variety of factors that have contributed to low uptake of this evidence-based lifestyle change program. However, the role neighborhood socioeconomic disadvantage plays in NDPP access and use has been underexplored. The state of Maryland is an ideal setting to investigate how neighborhood socioeconomic disadvantage impacts various dimensions of NDPP access due to the significant investments in primary care transformation and NDPP. This dissertation examines: (1) the relationship between neighborhood socioeconomic disadvantage and potential access to the NDPP using primary care providers geographic proximity to the NDPP sites in Maryland, (2) the relationship between neighborhood socioeconomic disadvantage and potential access to the NDPP based on geographic proximity of individuals with prediabetes to the nearest NDPP site in Maryland, and (3) the relationship between neighborhood socioeconomic disadvantage and utilization of diabetes prevention intervention such as NDPP, metformin, or both among individuals with prediabetes in Maryland. Findings from this work underscore how targeted statewide public health and health care initiatives can enhance NDPP access and utilization in neighborhoods with higher levels of socioeconomic disadvantage.