The June 2012 Supreme Court decision on Medicaid expansion provides insight into how physicians and the general public engaged with the political and policy-making process. This study assesses the making of public statements on Medicaid policy, through analyzing letters to the editor published in one hundred newspapers across the United States. A total of 2,792 individual letters to the editor were coded for their stance, use of evidence, use of personal experience, and framing of the Medicaid expansion. Both the general public and physicians expressed support for expansion in their letters, but physicians were more likely to emphasize the role of Medicaid expansion at improving the quality of care and saving lives. Additionally, physicians were more likely to cite evidence and personal experiences while framing their position on Medicaid expansion. There is no evidence of a shift in stance on Medicaid expansion after the election of Donald Trump, and Medicaid policy changes made by the new administration highlight the importance of the public engaging in the political and policy-making process of Medicaid.
One of the distinguishing characteristics of the Affordable Care Act (ACA) is that, while reform was national in scope, the implementation of much of the law rested on state involvement. In the case of the Medicaid expansion, state participation was the unexpected outcome of a judicial decision. One of the central goals of the ACA was the extension of health insurance coverage, and the Medicaid program was one of the foundations of increased insurance coverage (Frean, Gruber, and Sommers 2017). As written and passed by Congress, all states would be required to expand their Medicaid programs or risk losing federal funding for the health program. However, in 2012 the Supreme Court held that no states could be compelled to expand Medicaid (National Federation of Independent Business et al. v. Sebelius, 567 U.S. 519 ). While the ACA was still the law of the land, the flexibility given to states in the wake of the Supreme Court decision provided an opportunity to understand how physicians and the general public engaged in the political and policy-making process of Medicaid expansion.
There are many ways in which an individual can participate in the policy-making or political process, including voting, donating money, or volunteering with a campaign. However, in each of these instances, the driving force behind participation is not always clear. For example, vote choice by an individual is not always predicated on a single issue, such as support for Medicaid expansion, but is a product of a series of interconnected social, economic, and personal factors (Beck et al. 2002; Nickerson 2008). However, the making of public statements in support of or opposition to a particular policy can be analyzed and provides insight into the explicit choice of how individuals frame and support their stance on a particular policy.
This study sought to assess one form of public statement, letters to the editor, to understand how physicians and the general public responded to and expressed their opinion in response to the flexibility to expand Medicaid. A total of 2,792 individual letters to the editor from one hundred newspapers were collected, coded, and analyzed to understand the framing of Medicaid expansion and how authors used evidence and their own personal experiences in making their statements. From this larger pool of letters to the editor, physician authors were identified and their statements were analyzed to understand their perspective and engagement in the political and policy-making process, as well as differences from the general public and within the medical community.
Understanding the stance and framing of physician-authored letters to the editor are particularly interesting because they are consistently ranked as the most trusted source of views on health policy (Newport 2002), the most likely to “recommend the right thing for reforming the U.S. healthcare system” (Saad 2009: n.p.), and physicians consider themselves to be well informed on the ACA (Ganjian et al. 2015). President Obama sought to leverage the role of physicians during the health reform debate by presenting an outline of the White House's principles for health reform in front the American Medical Association annual meeting (Obama 2009b), as well as encouraging physicians in a Rose Garden speech “to fan out across the country and make the case about why this reform effort is so desperately needed. . . . Nobody has more credibility with the American people on this issue than you do” (Obama 2009a).
Newspapers provide one venue for physicians and the general public to make their personal case for or against health reform, express their stance, and engage in the political and policy-making process. Letters to the editor aid democratic norms of participatory democracy by providing an outlet for a range of viewpoints (McCluskey and Hmielowski 2012; Hoffman and Slater 2007; Wahl-Jorgensen 2001). The publication of letters to the editor allows the general public to shape debate, declare an individual or group stance, and raise awareness around a particular policy (Hynds 1992; Perrin and Vaisey 2008; Richardson and Franklin 2004). Even as rates of circulation and revenue have declined over time (Barthel 2017), newspapers remain an important provider of much of the public's knowledge about health and health policy (Brodie et al. 2003; Barabas and Jerit 2009; Ries, Rachul, and Caulfield 2011), and local news organizations, including the newspapers in this analysis, are consistently ranked as the most trustworthy source of information (Pew Research Center 2017).
This study of framing Medicaid expansion found that there was widespread support for the policy in published letters to the editor, by both physicians and the general public. Additionally, there is remarkable similarity in the choice of frames selected to support their stance toward the Medicaid expansion. However, there is a significant difference in the use of evidence and personal experiences as part of these public statements, with physicians much more likely to use both types of supportive material in their published letters. There was no evidence of any shifts in stance toward Medicaid after the election of Donald Trump, but changes in Medicaid policy under the new administration highlight the importance of the public engaging in the political and policy-making process of Medicaid.
Role of Physicians in Political and Policy-Making Processes
The role of physicians in the political and policy-making process in the United States has undergone a dramatic shift. As described by Paul Starr (2008: 4), the medical community has been transformed from disorganization and a position of relative weakness to a “position of cultural authority, economic power, and political influence.” Physicians have become a bloc of professionals that direct the “policy and lay perceptions of health problems” (Freidson 1994: 31) and have developed the trust of the general public to act in the best care of patients, which they have used to increase their own political power (Peterson 2001).
Physician professional organizations have supported and encouraged the active involvement of physicians in the political and policy-making process. For example, the largest professional organization for physicians, the American Medical Association, includes the statement that “physicians, individually and collectively through their professional organizations and institutions, should participate in the political process as advocates for patients” (American Medical Association 2016: n.p.) in its Code of Medical Ethics. Additionally, the American College of Physicians, the largest medical-specialty organization, encourages members to lobby and “educate the public, patient groups, and policymakers” (Snyder 2012: 91). Similar views have been issued by other physician specialty organizations (American Academy of Pediatrics 2018; American Psychiatric Association 2018).
Prior research has shown a variety of ways in which physicians have engaged, or not, in the political and policy-making process. Although a high proportion of physicians report being registered to vote (Antiel et al. 2014), there is little evidence that they participate in this basic form of political engagement. Compared to the general public and with other well-educated occupations, physicians were significantly less likely to vote (Grande, Asch, and Armstrong 2007). Additionally, while 91% of physicians rate political involvement as important, only a quarter of physicians reported completing any activities related to nonvoting political involvement (Gruen, Campbell, and Blumenthal 2006). While some physicians have reported being engaged in some aspect of the policy-making process, including collaborating with advocacy organizations, providing advice to government officials, and conducting policy related research, there is significant variation across rank, age, gender, and other demographic characteristics within the medical community (Jacobs, Greene, and Bindman 2013).
The likelihood of engaging in the political and policy-making process is not equal across the general public. Rather, the socioeconomic and educational background of individuals can inform whether they choose to engage in “influencing the selection of governmental personnel and/or the actions they take” (Verba, Nie, and Kim 1987: 1; see also Verba, Schlozman, and Brady 1995). For example, only an estimated 14% of the general public have authored a letter to the editor that was published by a newspaper (Zukin et al. 2006). Of those who do author a letter, they are more likely to be wealthier, more educated, and generally more engaged in the political process (Perrin 2016). Physicians, with their education, prestige, and power, are an important segment of the population to shape opinion and policy. However, just as with the general public, there is variation in opinion on the Medicaid expansion policy within the medical community. The political beliefs of physicians influence how they practice medicine (Hersh and Goldenberg 2016) and shape their stance on policies (Pollack, Armstrong, and Grande 2017). This variation is particularly marked by the Medicaid expansion policy, which is entwined with the broader polarization of health reform, as well as differential importance of expansion across medical occupations. Medicaid expansion as a policy is more visible to physicians and beneficial to patients among particular medical specializations. For example, for emergency department physicians, Medicaid expansion provides opportunities for increasing the quality and consistency of medical coverage for the previously uninsured patients going to hospitals to receive care. Conversely, due to low reimbursement rates, some physicians choose to not accept Medicaid patients in their practices (Hing, Decker, and Jamoom 2015), and Medicaid expansion as a health policy does not affect their business practices.
Use of Evidence and Personal Experience
Authors can supplement their stance and framing of the Medicaid expansion policy by leveraging evidence and personal experience. Studying Medicaid expansion as a policy will help us understand how physicians and the general public engage with and use prior evidence to support their stances. A growing body of evidence touches on a variety of different effects of the Medicaid expansion, including research on the role of Medicaid expansion on health improvements for the newly insured (Jacobs, Duchovny, and Lipton 2016; Miller and Wherry 2017; Wen, Druss, and Cummings 2015), improving state budgets and finances (Price and Eibner 2013; Buettgens, Holahan, and Recht 2015; Blumenthal and Collins 2014), improved financial well-being of the newly eligible (Hu et al. 2016; Brown, Kowalski, and Lurie 2015), and as an economic stimulus for hospitals and small businesses (Nikpay, Buchmueller, and Levy 2016; Dranove, Garthwaite, and Ody 2016; Graves 2012). Not only the peer-reviewed literature but also think-tanks across the political spectrum have examined the effects of Medicaid expansion.
While there has been a growing emphasis on evidence-based policy making (Gray 2004; Oliver, Lorenc, and Innvær 2014), there is still little use of evidence in the policy-making process, particularly among elected policy makers and political elites (Cairney and Oliver 2017; Grogan, Singer, and Jones 2016). There has been less of a focus on the use of evidence among the general public, particularly in understanding the diffusion of evidence and how the general public may or may not engage with health services and health policy research to support their position on the Medicaid expansion. Physicians, by their training and education, are likely to be more comfortable leveraging evidence that will support their stance toward Medicaid expansion. Even though the general public can access the same literature and research, their unfamiliarity with the evidence can dissuade them from using this supporting material as part of their stance toward Medicaid expansion.
Additionally, letters to the editor provide authors the opportunity to use their own personal experiences to support their stance toward the policy. Unlike with the use of evidence, individual authors do not need any special training, education, or familiarity to leverage their own experiences. The general public and physicians can draw from their positive or negative experiences with the Medicaid program to justify their stance toward expansion.
Newspaper Sampling Strategy
Use of newspapers and letters to the editor to analyze framing of a public policy has been employed by earlier studies (Cooper, Knotts, and Haspel 2009; Racine et al. 2010; Perrin 2016). There are a variety of potential selection criteria for newspaper inclusion. Prior work on framing and communication has focused primarily on a smaller number of newspapers with large circulations (Barry et al. 2011) or an in-depth study of several newspapers in a specific geographic location (Moreland-Russell et al. 2012). Sampling from the largest circulation newspapers (e.g., New York Times, Wall Street Journal, and Washington Post) or only within a geographic location has drawbacks. It does not account for the state-specific circumstances that encompass decisions to implement the Medicaid expansion occurring at the state level, or the variation in different frames used across states. While national newspapers can provide an account of the broader trends related to health reform, they miss the nuance of state-specific letters published in newspapers that are influential to state policy makers.
To understand how physicians and the general public engage in the political and policy-making process by writing letters to the editor, two newspapers were selected from each state, mirroring earlier studies (Gollust and Lantz 2009; Gollust, Niederdeppe, and Barry 2013; Hoffman and Slater 2007; Rose 2015). Selection of newspapers into the sample followed a twofold process. First, the two most populous Metropolitan Statistical Areas (MSAs) as designated by the US Census Bureau, were identified in each state. Second, using data on newspaper circulation, the newspaper with the highest circulation in each of the selected MSAs was then chosen for inclusion in the analyses (Alliance for Audited Media 2018). This process resulted in one hundred newspapers included in the sample, with two from each state. (See appendix A for full list of newspapers included.) One deviation from the selection of newspapers should be noted: Rhode Island and Hawaii are comprised entirely of one MSA, so the two highest circulation newspapers in these states were selected for inclusion instead.
The collection of letters to the editor in each newspaper used Newsbank and Gannett Newsstand databases. A hybrid standardized search criteria was used for each newspaper to find and collect letters to the editor. The same keyword (Medicaid) and date range (March 23, 2010, through March 31, 2017) was used for every newspaper. However, one component of the search criteria varied across each newspaper: the specific section of the newspaper the term Medicaid would be searched within. Each newspaper used a slightly different designation for this section, which was used in conjunction with Medicaid to identify the letters to the editor. For example, in some newspapers the letters to the editor were in the Opinion/Editorial section; in others, the op-ed section. Detecting the appropriate section to include in the search criteria followed an iterative process. First, using the URL of the selected newspaper, a recent letter to the editor would be found related to any subject. Then, using the appropriate Newsbank or Gannett database, that randomly selected letter to the editor would be found, which would then indicate the appropriate additional search term to include with Medicaid. This same process was followed for each of the one hundred selected newspapers.
Coding the Letters to the Editor
Each letter to the editor was hand coded. Prior to coding the letters, coding frames were developed from a combination of provisional and open coding techniques (Strauss and Corbin 1998). The initial coding themes were developed through previous work related to framing the rhetoric of the Medicaid expansion, as well as generated from prior literature. A subset of one hundred letters was initially reviewed, including a secondary coder, to determine the effectiveness of the coding instrument, and the two coders agreed on 78% of the coding. From this subset analysis, the coding protocol underwent further development to ensure consistency in the coding process.
As each of the letters was hand coded, the list of different frames was expanded to account for nuances in the arguments made in support and opposition to the Medicaid expansion. There was no limit on the number of codes that could be given to each letter. At the end of the coding process, the different frames underwent axial coding, where codes related but distinct were subsumed into broader categories (Miles, Huberman, and Saldana 2013). In total, the analysis considers eleven different types of codes. (See appendix B for examples of codes included in the final analysis.)
In addition to coding the different arguments and frames, several other variables were collected for each letter. First, each letter was judged on the overall stance of the letter writer (positive or negative) on the Medicaid expansion. Second, the author's use of evidence and personal experience was collected for each letter. For evidence, the author of a letter needed to express a specific source for data that they presented. For example, if a letter reported that “a recent Harvard medical study estimates between 7,000 and 17,000 U.S. citizens will die each year because they lack the basic care that would have been provided them by Medicaid expansion through Obamacare” (Daniel McMeen, Wisconsin State Journal, November 30, 2014) then it was coded as demonstrating use of evidence in their letter. Conversely, if in a letter the author wrote, for example, that “28 states and the District of Columbia have now taken advantage of the chance to expand coverage using federal dollars, and many are actually saving money as a result” (Jan Moller, Advocate [Baton Rouge, LA], July 30, 2015) but did not provide any source for their claim of saving money, the letter was coded as not using evidence. Additionally, the use of personal experience was collected for each letter. For example, if a letter said, “As a social worker, I have a hard time accepting that our state wants to turn down the federal funding for Medicaid expansion. . . . They are people like my former client, a blue-collar worker who couldn't afford to get a physical because all her money went to family and rent. Her cervical cancer went undiagnosed for far too long” (Carla Damron, State [Columbia, SC], February 12, 2013), to describe support for expansion, then it would be coded as using personal experiences.
The last piece of information collected from each letter to the editor was to identify if the author was a physician, involving three different criteria. First, authors who are physicians self-identified in 87% of the letters ultimately included in the population of letters written by physicians, either in the text or in the signature line of the letter. Second, the first name, last name, city, and state of each author of a letter to the editor was compared against National Plan and Provider Enumeration Systems records for a National Provider Identifier (NPI; npiregistry.cms.hhs.gov/). (Licensed health care providers use NPI numbers to help facilitate provider payment.) Third, because the NPI database is only for a point in time, each author's full name and state of residence was searched through an Internet database to determine if there were any previous records of a physician with that name practicing medicine. For all individuals who were identified as a physician, the specialty was also included in the analysis.
Logistic Model Specification
To understand differences in the types of frames, use of evidence, and personal experience across the letters to the editor, the letters were divided into two populations for logistic regression. The first set of logistic regression models contained all 2,792 letters to the editor. The second set of letters comprised only those written by a physician. The same logistic regression models were run for both sets.
The analysis used three main logistic regression models each with a different binary outcome variable of interest—stance, use of evidence, and personal experience—plus a series of control variables: specialty of physicians, divided into primary care (family medicine, pediatrics, internist, geriatrician, and obstetrician/gynecologist) and specialists (pulmonology, psychiatry, nephrology, ENT, surgeon, oncology, cardiology, urology, anesthesiologist, pathology, ophthalmology, neurology, and dermatology); implementation status of the state where the author resided at the time of publication; and year the letter was published, to understand if framing, stance, or use of evidence and personal experience changed over time, with special designations for all letters written prior to the Supreme Court decision in June 2012 and for all letters published after the November 8, 2016, election of President Donald Trump.
Table 1 shows the percentage breakdown of specific frames used by the authors of letters to the editor. (See appendix B for examples of each frame.) In total, 2,792 different letters to the editor were collected. Of those letters, 295 (11%) were identified as being written by a physician, with the remaining 2,497 letters written by the general public. The physician authors of letters to the editor were divided into several different specialties; the most common specialty was internal medicine (n = 82), followed by family medicine (n = 63) and pediatrics (n = 34). Overall, 197 letters were from individuals identified as a primary care physician, with the remaining 98 written by specialists.
In total, 708 different frames were used by physicians, and 5,172 by the general public (multiple frames were employed in a single letter). The most frequently used frame for both physicians and the general public was the opportunity for a state to improve access to health services by expanding Medicaid, representing 21% of the total frames in both groups. For physicians, this was followed by the opportunity to improve mortality for the newly insured; for the general public, the second most common frame was the financial benefits of expansion. Adverse financial ramifications of expansion was the most commonly negative frame among physicians and the general public, representing 4% and 7% of the total, respectively. Overall, the distribution of frames between the general public and physicians was quite similar; the only framing categories with statistically significant differences between physicians and the general public was the physicians' more frequent emphasis on improvements for mortality and quality of care if expansion is implemented and the general public more frequently invoking politics to frame their stance.
There are three key results from the logistic regression analysis of letters to the editor. The first is the strong support for Medicaid expansion in general, across both physicians (83%) and the general public (76%; figure 1). Yet, even with support across the total populations of letter writers, physicians were more likely to express support for the policy than the general public (χ2 = 5.9, p < 0.01). Among the physician letter authors, there was a statistically significant difference between specialist and primary care physicians (table 2): a specialist was 73% less likely than a primary care physician to have a positive stance on expansion (p < 0.01).
In comparing the effect of time on stance, writers of letters before states could implement the Medicaid expansion in 2014 were much more likely to oppose the policy. This opposition diminished over time: once states were given the flexibility to implement the Medicaid expansion, the number of letters supporting the policy increased. This support for Medicaid expansion was not diminished by the election of Donald Trump in 2016, with his promises to “repeal and replace” the ACA: there was no evidence of any statistically significant shifts in stance, use of evidence, personal experiences, or types of frames used in letters published after the 2016 election.
Lastly, the use of evidence and personal experience was quite different between physicians and the general public: physicians cited evidence in 39% of letters, while the general public included evidence in 17%. A similar dynamic was evident in the use of personal experience: 37% of physicians and 11% of the general public referenced their personal life in their letters. Results from the logistic regression models for evidence and personal experience share similarities with the model on stance. In both cases, the general public was much less likely to invoke evidence or personal experience in support or opposition to the Medicaid expansion. There is also a positive combinatorial relationship between use of evidence and personal experience, as well as on the use of evidence or personal experience and stance: an individual who used evidence was 50% more likely to use a personal experience in the letter; an individual who took a positive stance toward the Medicaid expansion was nearly three times as likely to supply evidence or a personal experience in the letter.
Several limitations should be considered in conjunction with these results. First, there are additional avenues in which individuals can express their opinions, beyond the pages of the newspaper, including social media platforms, which may attenuate the need for individuals to write a formal letter to the editor to express their opinion on the policy-making or policy-implementing process within a state. Additionally, the gatekeeping procedures and editorial emphasis across different newspapers can influence both the number and the type of letters a newspaper will include in the printed news (Hogan 2006). Lastly, the use of NPI data to identify physician authors from the total population of letters could potentially misidentify physicians. While there were three different checks on each of the authors of letters to the editor to identify physicians, there are no mechanisms in place at the newspaper level to ensure that authors of a letter to the editor are who they purport to be. It is also possible that an individual who is not a physician but shares a name with a physician in the same city and state could be erroneously counted as a physician. Further research could analyze the role of membership in advocacy organizations and the types of frames selected by authors.
The results of this study have two implications. First, the absence of cited evidence and personal experience in public statements of support or opposition to the Medicaid expansion raises concerns over the broad diffusion of evidence around a policy. While this study uses a strict definition of evidence, overall it is more likely that authors of letters would not include any evidence or personal experiences to support their stance on the Medicaid expansion. However, there is a clear distinction between physicians and the general public in the use evidence to support positions. Merely leveraging evidence in a letter does not necessarily make for a better or more convincing public statement. Rather, it does highlight a higher level of familiarity and comfort with research among physicians, as well as higher than average levels of education than the general public. Physicians' use of evidence during the political and policy-making process, compared with the general public, is also demonstrated across different specialties: there was no statistically significant difference between primary care physicians and specialists in their use of evidence. Even as the proliferation and access to evidence about the Medicaid expansion has increased, these results indicate that the general public is less likely to use this evidence in their discussion of the Medicaid expansion policy.
Second, even with large differences in the use of evidence and personal experiences between physicians and the general public, overall there is remarkable similarity in how both groups framed their support or opposition to the Medicaid expansion. Letters were overwhelmingly positive in their stance toward the Medicaid expansion, and both groups of authors most frequently couched their stance on improving access for the newly eligible. These results highlight the narrow band of frames commonly leveraged by the letter authors. One frame category with significant differences between physicians and the general public, the improvement of mortality and quality of health care for the newly eligible, points to physicians' familiarity and comfort with framing their stance using medical research and examples. Additionally, physicians were less likely to engage in explicitly political arguments as part of their stance, compared with the general public. For a policy as politicized as health reform, physicians were much more comfortable than the general public in focusing on the medicine, science, and morality of expansion.
Even though there was no evidence of any shifts in stance or framing in the wake of the election of Donald Trump, the politics of health care and health reform will continue to be a contentious policy issue. The importance of public input and engaging in the political and policy-making process increased as congressional Republicans failed to “repeal and replace” the ACA. In the wake of this failure, the Trump administration's approach to Medicaid policy has focused on increased flexibility in modifications to the Medicaid program and devolution of administrative authority to the states. Going forward, as state policy makers continue to grapple with Medicaid policy during the Trump administration, there will be continued opportunities for physicians and the general public alike to engage with the political and policy-making process.
I thank Jared Gamm and Michael Hornstein for their assistance in data collection. I am also grateful for helpful comments and insights from Jeffrey M. Berry, Andrea Louise Campbell, and the other participants at the JHPPL conference held at Tufts University.