Background

Health practitioners, researchers, and advocates alike call for policies that influence health outcomes to be grounded in the best possible evidence. However, political scientists assert that evidence is only one of many important drivers of policy; political considerations, habit, anecdotes, stereotypes, and “gut instincts” may be equally or even more important to policymakers. This research explores the question of how policymakers use scientific evidence in making state-level policy decisions on two contested maternal and reproductive health issues in the U.S.: abortion, and the use of alcohol during pregnancy. Both these issues are current dynamic policy issues at the state level, affecting important aspects of maternal and reproductive health, with large numbers of restrictive or punitive state laws being passed in apparent contravention of current evidence. The aim of the work is to shed light on why state policies in these areas do not appear to reflect current best evidence on these issues, and explore implications for reproductive and maternal health practitioners and researchers.

Methods

Data for the study are drawn from semi-structured interviews with 29 state lawmakers and their aides in Maryland, North Carolina, and Virginia. These neighboring states share many socio-political similarities yet have different mixes of policies on the two health areas of interest. Participants were recruited from the primary health-related committees of the General Assembly in all three states, as well as from among sponsors and co-sponsors of 2017 bills on abortion or substance use in pregnancy, and members of committees that voted on 2017 bills on abortion or substance use in pregnancy. Interviews were conducted between March and July 2017. Interviews were audio-recorded, transcribed, and uploaded into Dedoose, qualitative data analysis software, for coding and analysis by inductive and deductive methods.

Findings

Papers One and Two present findings from the inductive analysis regarding use of evidence in abortion policy (Paper One) and substance use in pregnancy policy (Paper Two). We find that evidence does not instrumentally shape state legislators’ views on these issues. Legislators appear to trust anecdotes more than evidence; they feel that the knowledge they get from specific personal anecdotes is more “real” than scientific evidence. We find that despite evidence pointing to the harms of alcohol use in pregnancy, most lawmakers were not concerned about this topic; instead they prioritize the issue of opioid use in pregnancy over that of alcohol in pregnancy. However, evidence does appear to inform legislators’ high-level understanding of some issues, particularly regarding the safety of abortion.

Paper Three presents the results of a deductive analysis, applying a conceptual framework on policy use of evidence from political science literature to data from the current study. This paper reinforces prior political science literature that state lawmakers may use evidence more to substantiate and support existing policy preferences than for any other use, even “citing” “evidence” that doesn’t actually exist. However, legislators widely refer to and use evidence symbolically in a way that may help reinforce the value of evidence in policy deliberations.

Conclusions

While our study finds that evidence does not directly inform policy decisions, we also find that evidence plays an important role in policy making. Even if it is not instrumental to decision making, research is still important to motivate, inform, and even provide persuasive “ammunition” to those lawmakers who are predisposed to support a given position. This work has important implications for the fields of reproductive and maternal health. If practitioners and researchers grapple with the political reality of use of evidence, this may help them target their research dissemination efforts more strategically and effectively. Such a pragmatic approach to policy use of evidence may provide the best hope for good research to be applied in improving reproductive and maternal health.

In 2005, Ethiopia enacted a rare reform, liberalizing abortion law as part of the overhaul of its Penal Code. Unsafe abortion is one of the three leading causes of maternal mortality in low-income countries. However, few countries have reformed their laws to permit safer, legal abortion, and medical associations typically avoid reform advocacy. The three studies presented in this dissertation examine the political origins and theoretical implications of this unusual reform in Ethiopia, as well as the roles of health providers and their medical societies in reform and implementation. The first study describes the political history of the reform and assesses the utility of theories of agenda setting. The second study explores why, counter to theoretical expectations and empirical accounts, a medical society actively supported reform of national law on abortion. Finally, as laws in resource-poor countries are often weakly implemented, the last study looks at midwives' roles in implementation by assessing what underlies their decisions about abortion service provision. Data for the three studies are drawn primarily from fifty-four interviews conducted in 2012 with obstetrician-gynecologists and governmental and non-governmental actors involved in, or familiar with, Ethiopia's reform. The other data sources were a 2013 survey of midwives (n=188) at a national professional meeting and 12 interviews with third-year midwifery students. The first study finds that the liberalization of the law on abortion was facilitated by the ruling party and its receptiveness to reform, civil society actors collaborating closely with government, the "open windows" offered by the vehicle of the Penal Code reform and the momentum of reforms to improve women's status. This suggests that agenda setting theories focusing on national rather than international actors conform better to Ethiopia's reform process. The unusual medical society participation in the reform process was motivated by obstetrician-gynecologists' personal and organizational commitments to reducing maternal mortality as well as organizational and political circumstances that relaxed or removed constraints on policy involvement. The study of midwives' attitudes about implementing reform found that most (72%) were willing or possibly willing to provide abortion services, and that this willingness was positively related to attitudes toward abortion, knowledge of the law and past provision of medication abortion. Willingness to provide services was negatively related to their number of years as a midwife and religiosity. Interview data suggest complex dynamics underlying these results, including conflict between strong professional norms and personal religious beliefs. These findings can deepen our knowledge of policy-making processes in Sub-Saharan Africa and give insight into how medical professionals participate and understand their roles in policy making and in implementing controversial health policies.

In Mexico, as in many other countries, HIV/AIDS strategies are developed at the federal level and implemented at the state level. Local programs are expected to use data, in particular surveillance data, to drive their decisions on programmatic activities and prioritize populations with which the program will engage. Since the early 1980s Mexico has engaged a complex and consistent strategy to address HIV and as a result, a great deal of data is available to help target prevention efforts. However, data utilization varies by state due to a number of factors that influence the process. This dissertation uses the case of state HIV/AIDS programs in Mexico to identify and explore the factors that influence data utilization and evaluate their effect on these programs. Three distinct papers are used to explore this topic.

The first paper is a literature review that provides an overview of factors that have been previously identified to influence data and research uptake by decision-makers. These factors are sorted into three categories--macro political, resource and data characteristics--to develop a model for influencing decisions. This model is then superimposed over the cycle for surveillance data in order to clearly ascertain the effect these factors have on data utilization. In the second paper the model developed in the first paper is evaluated through a study in four state programs in Mexico. Interviews and a survey were conducted to assess how the influence factors facilitate and impede data utilization at the local level. Issues around communication, decision-making power, budgeting, data quality and dissemination were the primary concerns identified. The third and final paper explores decision space in these state programs. Specifically, this paper reviews the negative side effects that enforcing an existing policy against sharing antiretroviral drugs between states has had on the decision space of programs at the local level.

The results from this research show that in addition to the expected barriers to using data for decision-making, there are also a variety of subtle forces affecting local actors that also need to be taken into account if data utilization is to be improved. One of the main contributions of this study is the approach of studying macro political, resource and data factors simultaneously in order to assess the combined effect they have on decision-makers. This study also helps identify potential areas that both local and national-level actors can leverage to move towards data-driven programming, such as improving local capacity for data analysis and strategic changes to current data dissemination practices. Future research directions include additional comparative studies in Mexico as well as in other middle- and lower-income countries, further explorations of constrained decision space and how it affects program performance and an evaluation of the relative value of experiential knowledge in settings were data are absent or unreliable.

Early public health efforts in the United States addressed social conditions that contributed to poor health, with public health workers playing a role in large scale societal reforms, such as passing housing and sanitation laws, which led to diminishing deaths from infectious diseases. As chronic diseases became leading causes of death, public health research and practice became more focused on individual behaviors, widely thought to be the primary cause of chronic diseases. However, health inequities along the lines of place, race, class, and other forms or marginalization are still prevalent. A substantial body of research illustrates how social, political, economic, and environmental factors affect multiple health outcomes, including chronic diseases, and contribute to health inequities.

In public health practice, some local health departments (LHDs) and organizations that support their work have called for broader public health interventions that address social policies that contribute to health inequities in addition to providing direct services to individuals. With continued research and support, the approximately 2,800 LHDs nationwide can play a central role in reducing health inequities. However, engaging in this complex work necessitates new approaches, skills, frameworks, and organizational infrastructures for LHDs. The recent foreclosure crisis, which stands to increase racial and health inequities, provides a lens to examine whether and how LHDs can move from a rhetorical commitment to addressing social determinants of health (SDH) into actual public health interventions that reduce health inequities.

Through this dissertation, I examine LHDs’ role in the foreclosure crisis through three related papers. My aim is provide insight into how LHDs responded to the deep and fundamental shifts in access to stable and quality housing and wealth created by the foreclosure crisis that disproportionately impacted African-American, Latino, and some Asian/Pacific Islander communities. Through all three papers, I incorporate a focus on challenges and approaches to addressing the racialized causes and outcomes of the foreclosure crisis. My overall aim is to help advance local public health practice within LHDs to more effectively target the causes of health inequities, including gaining a better understanding of LHD approaches and needs related to addressing SDH through local policy.

In the first paper, A National Survey on Local Health Department Engagement in Addressing the Foreclosure Crisis, I describe the results of a national survey on LHD engagement in the foreclosure crisis, which includes LHD approaches to addressing foreclosure and barriers to engagement. Responses followed a diffusion of innovation pattern, with innovator, early adopter, early majority, late majority, and lagging LHDs. Respondents expressed a high level of interest in adopting innovative approaches to addressing SDH and described a need for models of how other LHDs are preventing or mitigating the impacts of foreclosure, especially through local policies.

In the second paper, Adopting an Innovative Public Health Practice to Address Foreclosure: A Case Study of Alameda County Public Health Department, and the third paper, Policy Entrepreneurs, Agenda-Setting, and Communication: An Exploration of How a Local Health Department Engaged in Addressing the Foreclosure Crisis, I describe findings from qualitative interviews with current and former ACPHD staff and partners. In the second paper, I identify factors that 1) differentiate ACPHD’s innovative approach from traditional LHD activities; and, 2) contributed to ACPHD being an innovator among LHDs.

Finally, in the third paper, I focus on ACPHD’s role as a policy entrepreneur in agenda-setting, including their communication approach. While the second paper focuses on how ACPHD developed into an innovative LHD in the area of local housing policy, the 3rd paper focuses on how in this role, ACPHD interacted in the local policymaking process. This case study also examines how the role of policy entrepreneur can be shared across two organizations (ACPHD and Causa Justa::Just Cause) and provides another way to conceive of entrepreneurism.

This dissertation examines an unlikely set of air pollution controversies in rural west Texas to illustrate the function of the fixing or shifting of scale to gain advantage or build support in environmental conflicts, and to explore possibilities for pragmatic environmental collaboration and problem solving using alternative scalar approaches. Such alternative approaches include, for government environmental agencies, "jumping" higher and lower within a hierarchy of spatial scales in order to match environmental goals, institutional capacities, and political realities; and for activists, forming networked alliances spanning multiple places, in order to scale up local efforts and engage with regulatory agencies and other "big" interests on a more equal footing.

The project brings together two theoretical streams: analysis of the politics of spatial scale, primarily occurring in the geographical disciplines; and theories of problem definition and agenda setting from the political science and public policy realms. Scale is a particularly apt concept within which to conduct this analysis, as it is a primary defining factor not only of the atmospheric phenomena of concern, but also the landscape, human geography and policy venues within which this case study takes place.

The case study begins with a controversy over visible haze pollution in Big Bend National Park in the early 1990s. Staff in U.S. and Texas environmental regulatory agencies initially attributed responsibility for the sulfur dioxide emissions causing the haze to two large coal-fired power plants in the city of Piedras Negras, which together constituted the major emissions source closest to the park. Strong opposition to this attribution of blame by Mexican government representatives opened the possibility of a jump from the "border" scale to the much larger domain of "regional haze." Subsequent technical analysis effectively shifted the scale of the haze phenomenon to a continental level, showing the polluting particles were emitted from locations far beyond the border. My proposal here is that when a scale jump occurs, the "old" scale does not disappear, but rather continues to frame social phenomena; this suggests a framework consisting of layered spatial scales. This assertion opens the possibility for an adaptive model of scale shifts, which presents the opportunity to engage environmental problems at different scales over time. In the present case, this means that scale jumps are not irrevocable; actors can choose to "try" a scale more than once as the problem's definition and context continue to evolve.

Meanwhile, environmental activists first emerged as an organized movement within the Big Bend in 1996, taking haze pollution-and its impacts on the region's spectacular landscape and tourist economy-as their founding issue. In subsequent years, local residents organized to around additional air pollution-related issues, including the siting of a new rock crushing plant. When confronted with these issues, community members conceived that problem primarily as a threat to their fiercely loved home and landscape. This home area had very specific geographical boundaries, with air pollution conceived as a particularly egregious breach of those boundaries-and pollution outside of those boundaries considered marginally more acceptable. However, I propose that activists' discourses on vulnerability, clean energy and border-area cooperation present an opportunity to form connections with broader activist communities across different locales, transcending the hierarchical state structures of environmental regulation. This type of network building strategy may provide an opportunity for local movements to scale up their efforts and engage more effectively with regulatory agencies, either as partners or as opposing interests.

Finally, I examine scale as one dimension along which science and place-based activism can engage with one another, either through cooperation or conflict. Public agencies hold promise as mediators for this engagement, with the ability to both lead and follow science across scales, take actions and monitor results over time, and yet maintain links to specific geographical jurisdictions and an imperative to democracy.

Abstract

Governor Decision-making: Expansion of Medicaid under the Affordable Care Act

By

Robin Flagg

Doctor of Philosophy in Health Services and Policy Analysis

University of California, Berkeley

Professor Ann C. Keller, Chair

This is a study of factors that influence gubernatorial decision making. In particular, I ask why some governors decided to expand Medicaid under the Accountable Care Act (ACA) while others opted against it. Governors, like all chief executives, are subject to cross-pressures that make their jobs challenging. Budgetary pressures may differ from personal ideology and administrative infrastructures may not allow for decisive moves. Add to the equation political pressures - in particular the pressure to align with partisan positions - and a governor is faced with a myriad of opposing and interrelated factors, each requiring attention, when taking a particular position. The calculation required of a governor when deciding upon a salient issue is thus extremely complicated and nuanced.

Although interesting in its own right, governor decision making is of additional significance because it may shed light on how the effects of increasing party strength and polarization are playing out at the state level. Partisan gridlock has dominated Congressional decision-making for much of the last decade. In Washington today partisan ideology dominates decision making. In particular, Republican elected officials increasingly espouse conservative policies and vociferously denounce any attempt at compromise. This study asks whether this ideology driven decision-making also exists at the state level. Specifically, I assess whether Governors are susceptible to the same partisan influences as elected officials in Washington and whether partisan politics and/or ideological polarization dominate governor decision-making as they do congressional actions. In particular, I study the factors involved in the decision-making process of each governor when deciding whether or not to support Medicaid expansion in his state. The focus of this study is the governor's calculations and considerations prior to "going public" with his position, irrespective of his success in getting his position adopted by the legislature. Specifically I explore the role the governor's party, the governor's personal ideology, the electoral results from the 2012 presidential election, the state's policy heritage, advocacy by state stakeholders, and the state economy played in the governor's calculations.

A mixed method research design is used, with each component (the quantitative and the qualitative) addressing a different level of question. The first part of this study is a descriptive and statistical quantitative review of all governors, assessing the various weights the 50 governors appear to give each of the seven factors discussed above. This in turn provides a context for the second part of the research, an in-depth case study providing a comprehensive analysis of how two governors made this politically salient decision

There are two main findings from this study: First, Mayhew's conclusion that congressmen legislate in a manner that promotes their reelection appears to apply in this case of governor decision-making. The partisan salience of the Medicaid expansion decision is a particularly strong test of this idea given that electoral pressures may influence less nationally prominent decisions, but weaken when partisan pressures are present and decisions are highly visible. And secondly, when studying variation across states, a mixed methods approach offers enhanced and nuanced findings as compared to a more quantitative model.

This study has found that many factors influence governors' decision-making. However, electoral pressure was not only the most significant in the general model but also appeared central to the case study portion of this research. While other factors (e.g., economics, existing institutions, the role of stakeholders, and the governor's religion) were found to be statistically significant in the general model, data from the qualitative portion of this research suggests that many of these factors may have played a role not in taking a position, but rather as justification for the position taken. In both cases, the underlying driver for the decision appears to be electoral interest: both governors studied were primarily concerned with ensuring that their decision on this highly salient issue was consistent with what they believed the majority of their electorate would support. As necessary, they used other factors to help frame their final decision in a manner that they believed would appease their electorate.

This study also highlights the power of a mixed method approach. While many of the findings of the general model are upheld by the case studies, without the rich information gleaned from the qualitative data augmenting the general model, the conclusions would have been too simplistic. The case study data portrays a number of examples in which the macro model over simplified the outcome and ultimately led to an incomplete or even erroneous conclusion. First, existing state political institutions (i.e., commissions, ballot initiative processes) and previous policy decisions render each governor's decision unique despite the fact that each maintains the same ultimate goal of political survival. Specifically, in Ohio the entire process of moving the decision out of the budgetary process and to the Controlling Board in Columbus was an attempt to avoid a polarizing vote and to allow otherwise ideologically opposed legislators to remain silent on expansion and ultimately retain party cohesion. This was clearly spearheaded by party leaders in support of their governor and instead of the internal party division projected by the general model.

Overall, this study affirms that small "d" democracy is alive and well. Because governors, like congressmen, are profoundly concerned with how their position presents to their constituents and thus their political futures, they ensure that their position on salient and visibly issues is either consistent with that of their voters or at least can be explained to their voters in a manner that neutralizes any divergence from the majority position. In the end, all politics is local and politicians must maneuver a frame to address their situation. In order to accurately assess how a governor manages the sometimes opposing pressures of ideology and politics, an in-depth case study is called for.