PROQUEST DTG-Local Public Health Performance and its Impact on Population Health

Wallace, Megan (2017) PROQUEST DTG-Local Public Health Performance and its Impact on Population Health. Doctoral thesis, John Hopkins University.

[img] Text
Local_Public_Health_Performanc.pdf - Published Version
Restricted to Registered users only

Download (2MB)

Abstract

Despite decades long calls for increased public health infrastructure, local health departments are still often overlooked. With continued and varied health threats being constantly posed, these health departments, which are considered to be the frontlines of public health practice, are being tested daily. All the while, many questions remain about the level of performance in local health departments and the mechanisms by which these health departments may become more effective in providing public health services to their communities. In this dissertation, I first present an in-depth look at the current level of local health department performance on the Ten Essential Public Health Services. Using principal component analysis, we calculate health department specific scores for each of the essential services based on reported activities. With these scores, we explore the distribution of performance across services and the correlates of health department performance. We find that there is great variability on the performance of the essential services across health departments, but performance tends to be low overall. Total population served and health department per capita expenditures were both consistently associated with health department scores. Cluster analysis showed that there are six distinct score clusters, which can be defined by their overall performance level (high, moderate, or low performance) and whether they focus on policy and regulation or direct patient services. These findings suggest that health departments may be specializing their services, to be more policy or patient service based, given their available resources or their role in their state’s public health infrastructure. We next explore a supplemental method to the current county health ranking systems. Ranking systems, such as the Robert Wood Johnson County Health Rankings, rank counties from first to last using morbidity, mortality and quality of life metrics. However, they do not include any sociodemographic context into the rankings, which leads to rankings that highlight the social inequities of health. While this is a valuable visual for the general public, and policy makers, it gives local health departments, who are well aware of these inequities, little new information. To explore this further, we created county health rankings driven by factors other than sociodemographic characteristics, which would provide health departments with new information on how they may improve their population’s health. Using cluster analysis, we create county groupings based on the sociodemographic make-up of the county populations. Using the outcomes of obesity, smoking, and motor vehicle mortality, we compare nationwide county rankings to within cluster rankings. We find that the relative performance of many counties differs greatly if they are being compared to cluster peers instead of a nationwide ranking. These findings challenge the current practice of comparing localities with no consideration of their sociodemographic context, and propose that comparisons between similar populations may provide more opportunities for learning exchange. Finally, we explore the premise that the activities local health departments perform positively impact health behaviors and outcomes in the populations they serve. Using hierarchical models, we examine the associations between local health department characteristics, including their scores on the Essential Public Health Services, and the health outcomes of smoking, obesity, seatbelt use, and a self-reported rating of general health in individuals living in the populations being served. Increased scores on several of the Essential Public Health Services were significantly associated with decreased odds of poor health outcomes. Higher health department scores on “Inform, Educate, and Empower” and “Link to and Provide Care” were both significantly associated w ith a decreased odds of reporting poor or fair health. Increased scores on “Inform, Educate, and Empower” were also associated with decreased odds of smoking. Decreased odds of inconsistent seatbelt was significantly associated with higher scores on “Monitor Health Status” and “Enforce Laws and Regulations”. While these associations between are noteworthy, perhaps the more remarkable finding was the strong and consistent relationship of people living in state governed local health departments having higher odds of poor health outcomes and behaviors than their locally governed counterparts after adjusting for the individual characteristics of the populations served and other health department characteristics. As population health in the United States comes under increased scrutiny, the need for a strong public health infrastructure must be on the forefront of these conversations. In this dissertation, we attempt to highlight the importance of local health departments in this dialog. Future research, including robust longitudinal studies and validation of the scoring methodology used in this work, would further increase the understanding of how local health departments function and what steps can be taken to improve their performance. This work supports previous findings that local health departments make meaningful impacts on the populations they serve, but many of them fall short of the expected level of capacity. A finding that I hope policy makers will note and work toward remedying.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: public health; performance
Subjects: 600 Technology (Applied sciences) > 610 Medical sciences Medicine > 613 Promotion of health
Divisions: Perpustakaan
Depositing User: K Kristiarso
Date Deposited: 13 Jun 2022 04:25
Last Modified: 13 Jun 2022 04:25
URI: http://repository.uinsaizu.ac.id/id/eprint/13877

Actions (login required)

View Item View Item