Comments
Abstract
Introduction: Pandemic influenza or other crises causing mass respiratory failure could easily overwhelm current North American critical care capacity. This threat has generated large-scale federal, state, and local efforts to prepare for a public health disaster. Few, however, have systematically engaged the public regarding which values are most important in guiding decisions about how to allocate scarce healthcare resources during such crises.
Methods: The aims of this pilot study were (1) to test whether deliberative democratic methods could be used to promote engaged discussion about complex, ethically challenging healthcare-related policy issues and (2) to develop specific deliberative democratic procedures that could ultimately be used in a statewide process to inform a Maryland framework for allocating scarce healthcare resources during disasters. Using collaboratively developed focus group materials and multiple metrics for assessing outcomes, we held 5-hour pilot community meetings with a combined total of 68 community members in two locations in Maryland. The key outcomes used to assess the project were (1) the comprehensibility of the background materials and ethical principles, (2) the salience of the ethical principles, (3) the perceived usefulness of the discussions, (4) the degree to which participants’ opinions evolved as a result of the discussions, and (5) the quality of participant engagement.
Results: Most participants were thoughtful, reflective, and invested in this pilot policy-informing process. Throughout the pilot process, changes were made to background materials, the verbal introduction, and pre- and post-surveys. Importantly, by holding pilot meetings in two distinct communities (an affluent suburb and inner city neighborhood), we discerned that participants’ ethical reflections were framed in large part by their place-based life experiences.
Conclusion: This pilot process, coupled with extensive feedback from participants, yielded a refined methodology suitable for wider-scale use and underscored the need for involvement of diverse communities in a statewide engagement process on this critical policy issue.
Studies suggest that another public health disaster such as the 1918 influenza pandemic would require intensive care unit (ICU) and mechanical ventilation capacity that are orders of magnitude greater than what is currently available (1). Even a less severe disaster could overwhelm critical care capacity, given that smaller disasters have had dramatic impacts on North American ICUs (2) that consistently operate at or near capacity (3).
Several groups have published recommendations for allocating life-support measures during a public health emergency. The guidelines differ substantially regarding which ethical criteria would be determinative, whether some groups of patients should be categorically denied access to life support (exclusion criteria), and whether it is permissible to withdraw a resource from one patient to provide it to another who is more likely to benefit (4–6). Ethical principles that have been suggested in published guidelines include saving the most lives; saving the most life-years; and the life-cycle principle, which uses age as a criterion. To date, public engagement regarding which ethical principles should guide allocation decisions has been limited either to polling of samples of the public, feedback on an already developed allocation framework, or as part of a larger discussion of crisis standards of care (5–7). Few efforts have attempted to specifically assess the allocation-related values of a community following a period of informed discussion in advance of creating a framework.
Public discussion about allocation in emergencies is critical for the following reasons:
| 1. | The public will bear the consequences of these decisions, | ||||
| 2. | The public’s values are central to choosing among multiple ethically permissible approaches to allocating scarce life-sustaining resources. | ||||
| 3. | A successful allocation effort will require public trust and cooperation, both of which are more likely if the development process has been transparent and inclusive. | ||||
| 4. | Knowledge about public perspectives and moral points of reference better prepares health authorities to develop appropriate allocation guidance and to communicate well about potentially charged issues. | ||||
Moreover, advance discussion is essential because in-depth deliberations will not be feasible during a public health crisis.
In health policy and other arenas, community engagement has been performed using varying methodologies at different points in policy development. Both timing and methodology can influence public feedback on complex issues. On a few occasions, public input has been solicited in advance of policy development to allow the “voice of the people” to shape initial policy (8). To optimize conditions for early public input, we chose the Deliberative Democracy methodology. This methodology requires citizen participation in open and informed conversations to bring multiple perspectives into the discussion and base dialogue upon the best information available (9).
Our project was a collaboration between the Johns Hopkins Medicine Office of Emergency Management, the Johns Hopkins Berman Institute for Bioethics, and the Center for Health Security at the University of Pittsburgh Medical Center (UPMC), with content expertise provided by the Carnegie Mellon Center for Ethics and Policy, the Carnegie Mellon Program for Deliberative Democracy (PDD), and the UPMC Program on Ethics and Decision Making in Critical Illness. The hypothesis of this pilot study is that deliberative democratic methods may be used to promote engaged discussion about complex, ethically challenging, healthcare-related policy issues. We aimed to develop and pilot-test deliberative democratic procedures and produce background materials that could ultimately be used in a statewide engagement process to inform a Maryland framework for allocating scarce healthcare resources during disasters.
The pilot meetings engaged citizens of Baltimore City and neighboring Howard County, Maryland, in discussions about the allocation of scarce lifesaving healthcare resources during catastrophic disasters. The specific meeting procedures adopted for the pilot study have been fielded successfully in other potentially controversial, value-laden policy issues, such as regional tax sharing, economic redevelopment, and same-sex marriage (9). The background materials used were developed specifically for this project using an iterative process. The materials framed discussions in terms of a hypothetical severe influenza pandemic that had overwhelmed Maryland’s critical care capacity. This study was approved by Johns Hopkins Medicine’s Institutional Review Board X (Protocol No. NA_00070411). Informed consent was obtained from all participants.
Team members from Carnegie Mellon University and the UPMC Program on Ethics and Decision Making in Critical Illness, along with the Principal Investigator (PI), took the lead in proposing the ethical principles to be considered and drafting the project materials. The relevant literature was reviewed. The final selection and description of the principles, summarized in Table 1, were the result of discussions among all investigators. The draft materials were reviewed by team members for clarity and accuracy and by laypersons for comprehensibility. The final materials included a 10-page background document and a 1-page summary of the 2 questions to be discussed, as well as the ethical principles and the scenario intended to frame the discussion.
| Principle | Description |
|---|---|
| First-come, first-served | Patients are given ventilators in the order in which they are diagnosed until no ventilators remain. This is a continuation of standard practice. |
| Lottery | Patients are assigned a number. Healthcare providers select a random number and give the ventilator to the patient with that number. |
| Prioritize those most likely to survive | Patients with the best chance of surviving their current flu illness are given priority over those less likely to survive. |
| Prioritize those with the most years left | Patients who are likely to survive the most number of years are given priority. Underlying illnesses the patient already has are considered. |
| Life cycle or “fair innings” principle | Patients who have lived through the fewest stages of life, from youth to old age, are given priority. |
| Value to others in a pandemic | The patients most likely to have an immediate impact on society—first responders, clinicians, or vaccine workers—are given priority. |
Participants were recruited using existing partnerships with community-based organizations, including the Johns Hopkins Hospital Department of Community Health, the Historic East Baltimore Community Action Committee, and the Howard County Community Emergency Response Network. Community leaders gave potential participants a letter summarizing the study’s purpose and containing contact information for registration. Healthcare workers were excluded because of concerns that they might dominate or bias the dialogue. For the second forum, the Horizon Foundation, a Howard County organization, created a website to aid in registration of participants. A similar website was created to organize participants for future forums (10).
Participants were convened for 5 hours on a Saturday and were divided into groups of six to eight based on demographic criteria to ensure diversity. Each participant completed a pre-survey to ascertain his or her thoughts on scarce resource allocation before the deliberations. The survey was followed by time allowed for participants to review the background materials and a summary of the problem delivered orally by the PI.
In the first phase of the deliberation, groups discussed the question, “What should we do in situations where there are more patients needing ventilators than there are ventilators to use?” The discussions were framed around the six ethical principles (Table 1), and participants were invited to introduce additional ethical principles for consideration. In the meeting’s second phase, the groups discussed the question, “Should healthcare providers ever be allowed to remove a ventilator from one patient who needs it to survive and give it to another patient who also needs it to survive?” During the third phase, each group posed one or two consensus-derived questions to an expert panel. The panel responded to the questions with the entire forum.
In the final phase of the meeting, participants completed a post-survey. A subset of participants volunteered for in-depth exit interviews that focused on their understanding of the issues and on the forum process. The meeting schedule is summarized in Table 2.
| Time | Activity |
|---|---|
| 9:00 A.M. | Registration |
| 9:00–9:45 A.M. | Convene in small groups |
| 9:45–9:50 A.M. | Welcome and agenda review |
| 9:50–10:15 A.M. | Opening remarks |
| 10:15–11:15 A.M. | Phase I: Small-group discussion |
| • What should we do in situations where there are more patients who need ventilators than there are ventilators to use? | |
| 11:15 A.M.–12:00 P.M. | Small-group discussion |
| • Should healthcare providers ever be allowed to remove a ventilator from one patient who needs it to survive and give it to another who also needs it to survive? | |
| 12:00–12:15 P.M. | Phase II: Small-group discussion |
| • Tables develop two questions to ask the panel | |
| 12:15–12:30 P.M. | Break and pick up box lunches |
| 12:30–1:30 P.M. | Phase III: Working lunch and panel discussion |
| 1:30–1:50 P.M. | Phase IV: Small-group discussion |
| • Group reflections on panel and day’s discussion | |
| • Complete post-survey | |
| 1:50–2:00 P.M. | Closing comments |
The group facilitators were volunteers who had prior experience with facilitation (e.g., counseling), were graduate students in a related field, or were public health analysts. Notetakers were assigned to each group. The facilitators and notetakers attended a 3-hour training session run by the PDD to learn the methods used in deliberative democracy. Project management for the forums was provided by RESOLVE, a Washington, D.C.–based nonprofit facilitation organization. The PI (E.L.D.B.) and coinvestigators (including H.G., M.S.-S., R.C., T.D., P.B.T., A.R., and E.S.T.) participated in both meetings as either content experts or process observers.
The key outcomes used to assess the project were (1) the comprehensibility of the principles and background materials (i.e., whether participants understood the formal principles and the policy context), (2) the salience of the principles (i.e., whether the principles were meaningful and the significance they held for participants), (3) the perceived usefulness of the discussions (i.e., whether the participants saw the conversations as personally and socially beneficial), (4) the degree to which participants’ opinions evolved as a result of the discussions (i.e., whether the participants incorporated what they heard from others into their own comments), and (5) the quality of participants’ engagement (i.e., whether the participants were actively engaged in listening and/or speaking throughout the day).
All outcomes were assessed on the basis of the facilitators’ notes and debriefings (comprehensibility, salience, and quality of engagement), direct observation (comprehensibility and quality of engagement), pre- and post-surveys (comprehensibility, salience, evolution of opinions, and perceived usefulness), and exit interviews (comprehensibility, perceived usefulness, and evolution of opinions). An alternative to validation, triangulation (i.e., the cross-checking of outcome data obtained using different methods and from different observers) was used to assure the trustworthiness of the pilot study observations.
Pre- and post-surveys were drafted by experts in the PDD according to their standard protocols. The surveys were then refined by experts from the Berman Institute for Bioethics. After the first meeting, minor additional edits were made in keeping with the purpose of the pilot. Exit interview questions were formulated by investigators from the Berman Institute and revised in an iterative fashion with input from the entire project team. Direct observation and debriefings were conducted by investigators and RESOLVE staff according to RESOLVE’s standard facilitation and debriefing methods.
To evaluate the applicability of the process and materials in disparate contexts, the starkly different settings of inner-city Baltimore (specifically the East Baltimore neighborhood) and suburban Howard County, Maryland, were chosen for the meetings. The East Baltimore neighborhood struggles with high rates of crime and drug use (11). The median household income there is $18,522, and 28% of families live below the poverty line (12). A majority of residents ages 25 years or older (66.4%) have a high school degree or less (12). In Maryland’s state health rankings, Baltimore City ranks last (13). In contrast, Howard County is Maryland’s healthiest county, its residents have a median income of $105,692, and 4.5% of families live below the poverty line (14). A majority of residents ages 25 years and older (58.7%) have a bachelors’ degree or higher.
The first of two pilot meetings, held in East Baltimore in May 2012, had 39 participants, and the second, held 1 month later, involved 29 participants from Howard County. The characteristics of the groups are given in Table 3. Although age and sex distributions were similar in the two groups, participants in the East Baltimore group were more likely to be African-American and to self-identify with the Democratic Party. Our recruited samples tended to be older and more frequently female than the baseline demographics of the populations in the two areas on the basis of U.S. census data. In East Baltimore and Howard County, 67% and 66% of participants, respectively, were female, compared with 51% and 52% of all residents in the respective areas according to census data. Additionally, in East Baltimore, 41% of participants were older than 50 years of age, as compared with 34% of all residents in that neighborhood based on census data. The Howard County group showed a similar pattern, with 69% of participants being over 50 of age, as compared with 42% in that age group in the general population recorded in the census data.
| Demographic variable | Baltimore City | Howard County, Maryland |
|---|---|---|
| Mean age, yr | 50 | 55 |
| Females, % | 67 | 66 |
| Marital status | ||
| Married | 38 | 48 |
| Never married | 26 | 24 |
| Divorced/widowed | 28 | 14 |
| Race/ethnicity | ||
| Black | 67 | 21 |
| White | 15 | 52 |
| Hispanic | 10 | 0 |
| Political affiliation | ||
| Democratic Party, % | 82 | 59 |
| Income | ||
| Household income <$20,000/yr, % | 36 | 10 |
| Education | ||
| College degree, % | 49 | 86 |
Holding two pilot forums 1 month apart allowed the investigators to explore refinements of both the materials and the process in disparate settings. To support this refinement, the team assessed the participants’ experiences through (1) review of survey results (2), group facilitator debriefings, and (3) participant exit interviews. Revisions were made following group discussions if those revisions remained consistent with the deliberative methodology and all team members agreed.
Pre- and post-survey responses were compared for consistency, as was written feedback regarding the process and materials. Inconsistencies were noted in the responses to several survey questions, and revisions were made to enhance clarity (Table 4). Issues raised in team debriefings prompted changes both to the PI’s oral introduction and to facilitator training.
| Group Question | Response | Pre-Survey | Post-Survey | Change |
|---|---|---|---|---|
| East Baltimore: Do you think there are situations when healthcare providers should help one patient over another? | Yes | 73% | 54% | −21% |
| No | 6% | 26% | +20% | |
| Unsure | 19% | 20% | +1% | |
| Howard County: Are there certain people or groups of people who should receive first priority for treatment over others when there is not enough treatment available to help everyone? | Yes | 59% | 72% | +13% |
| No | 10% | 24% | +14% | |
| Unsure/no response | 31% | 3% | −28% |
In exit interviews, the East Baltimore participants expressed a desire both to understand the process of decision-making during a disaster and to have a voice in the decision-making. They wondered where our meeting fit in the process and how their input would be used. This concern was addressed in the Howard County meeting by expanding the explanation of the scope of the project. Both groups expressed concern that some demographic groups were underrepresented in the meetings, and efforts were made to revise the recruiting processes to minimize these issues, including plans for radio and newspaper advertisements to the general public for future sessions, in addition to recruitment through existing community groups.
Multiple methods were used in the overall assessment of each of the key outcomes outlined above. Specifically, both groups engaged vigorously with the ethical principles and actively participated in discussions, as assessed by both direct observation and facilitator debriefings held immediately following the sessions. Additionally, facilitator debriefings and participant exit interviews after both forums indicated that the participants understood the principles and the policy context (comprehensibility) and valued the process (usefulness). The survey results confirmed the perceived usefulness of the process, revealing that 97% of participants at both meetings felt that discussions such as the ones in this study are important. With regard to the salience of the principles, the post-survey indicated somewhat similar distributions of preferences regarding the principles in both forums (Table 5), but facilitators’ notes and direct observations indicated that the major areas of concern varied. This variation is discussed below. Finally, comparison of the pre- and post-survey results, both within and between the forums, indicated an evolution of thinking by many participants as a result of the deliberations (Figure 1), suggesting that the process was useful and valid in both settings.
| Principle | Never or Rarely | Sometimes | Always or Often |
|---|---|---|---|
| East Baltimore/Howard County | East Baltimore/Howard County | East Baltimore/Howard County | |
| First come, first served | 18%/52% | 45%/31% | 27%/13% |
| Lottery | 88%/80% | 9%/14% | 0%/3% |
| Give priority to those most likely to survive | 15%/6% | 42%/7% | 33%/83% |
| Give priority to those with the most years of life left | 39%/34% | 36%/38% | 21%/27% |
| Give priority to those who have lived fewer “life cycles” | 51%/38% | 24%/55% | 18%/3% |
| Value to others in pandemic | 18%/24% | 39%/21% | 33%/56% |
Figure 1. Pre-/post-survey comparison. Bars represent percentage of respondents on pre-/post-surveys at each location.
[More] [Minimize]In addition to the stated outcomes and metrics, conducting this pilot study in contrasting settings provided opportunities to explore the usability of the materials and the process in varied contexts. When the two meetings were compared, certain distinct concerns emerged that may be explained, in part, by the contrasting circumstances present in the two very different communities. People engage with medical ethics on the basis of their life experiences, social roles, political concerns, and cultural beliefs (15–17). People’s moral perspectives on medical practice emerge from what they have experienced and learned about the world, including relations of inequality, and not simply from abstract high-order values (16).
Residents of East Baltimore—a community beset by crime—spent a significant amount of time discussing whether particular groups, such as convicted felons, should be disqualified from access to ventilators. The question whether undocumented immigrants should have access to such resources also came up in East Baltimore, part of which has become home to an increasing population of Latino immigrants. Possibly because they were sensitized to the barriers that needy residents routinely face in accessing healthcare, the urban group felt adamantly that ability to pay should not determine access.
In East Baltimore, many participants emphasized that resource allocation decisions should not be implemented arbitrarily. The participants emphasized the importance of transparency and clarity with regard to both which criteria were being used and the rationale behind them. They expressed a desire to understand the process that would be used to develop Maryland’s framework for critical healthcare resource allocation during disasters, where the study’s forums fit into this process, and how their feedback would be influential. Such concerns would be well-founded among groups with historic reasons to question whether public policies represented the interests of disadvantaged people who are often politically marginalized.
Although social justice themes were more salient in the East Baltimore discussion, logistical concerns more often colored the discussions among participants in the suburban Howard County group. In general, the Howard County participants had more exposure to disaster planning than those in East Baltimore, having been recruited to the study by a private–public partnership focused on strengthening local resilience to disasters. The participants in this group struggled with the implementation time frame required by different approaches. Would someone die without a ventilator while decision-makers were still making up their minds? The Howard County residents also expressed greater concern than the East Baltimore residents about the potential for bias in decision-making. To ensure fair choices, the county participants recommended reliance on a rational mechanism—expressed as “science” or “statistical analysis”—to make allocation determinations.
Despite these differences, both groups understood the materials and the principles, as judged by the facilitators’ notes, direct observations, and exit interviews. The East Baltimore group was no less facile with the nuances of the conversation than their Howard County counterparts, who, on average, were more affluent and had attained more formal education.
Both groups felt strongly that no single principle could adequately balance the competing aims and values triggered by allocation decisions. Some felt that a combination of principles should be used. Both groups suggested alternative strategies, such as using organ transplant allocation criteria as a model or adopting a tiered approach by applying different principles at different stages in the process.
Among the East Baltimore group, although the “lottery” principle seemed fair in theory, participants felt that it would not be adequate in a pandemic setting. Similarly, the approaches labeled “prioritize those most likely to survive,” “prioritize those with the most years left,” and “life cycle” were felt to be morally relevant but inadequate on their own. Of note, the principle labeled “value to others in a pandemic” was considered the most problematic principle, with no clear consensus reached on its application.
Among the Howard County participants, the “first-come, first-served” principle was generally discarded as inadequate. Although the “equal chance” of the “lottery” principle appealed to some, others felt that it was “senseless.” This group found strengths and weaknesses with each of the other principles, but did not have a clear preference for any of them. Several participants expressed significant discomfort with the “life cycle” concept. Both groups emphasized the need, in the context of prioritizing younger patients, to ensure that enough adults were given access to ventilators so that they could survive to care for children.
When the discussion turned to removal of a ventilator, both groups offered widely varying opinions. Some felt that removing a ventilator amounted to killing one person to save another, thus making this option unacceptable. Others, however, suggested that “survivability” is “key.” Many who were comfortable with removing a ventilator from one person to give it to another insisted that the patient’s clinical trajectory be taken into account. Although most who commented acknowledged a possible need to reallocate ventilators, they were concerned about bias influencing those decisions.
Engagement of the public on healthcare-related topics that may be politically charged and ethically divisive poses significant challenges for both healthcare providers and policy makers. Consistent with our stated outcome measures in this pilot project, we have demonstrated that (1) deliberative democracy methods can be used to generate coherent public feedback that may ultimately be useful in formulating health policies, (2) materials generated for use in the deliberative process for this specific issue were comprehensible and useful to a diverse group of participants, and (3) deliberative democratic methods may reveal variations in approaches to health policy issues that are deeply rooted in specific communities and transcend the specific content of the discussion. This last finding reinforces the value of recruiting diverse participants to uncover the range of salient issues in a pluralistic society.
The deliberative democratic method provided a format for facilitating civil conversation that might otherwise be fraught with both misinformation and contentious debate. By giving participants meaningful background information and a framework for dialogue, we were able to elicit multiple opinions in conversations characterized by mutual respect. Many participants altered their views after listening to other participants’ rationales.
The participants’ engagement in the project was striking. They grasped the depth and breadth of the issues raised, and they indicated through their investment in the conversation that health policy issues are important to them. The apparent difference in the overriding concerns expressed by the two groups was also remarkable. Our preliminary findings—from only two sessions in demographically contrasting communities—suggest that there may be differences in the ways in which communities approach healthcare allocation issues in a disaster. Because we are confident that participants in both communities understood and were engaged with the issues and principles discussed, these differences are not likely to be a consequence of the deliberative democracy methods’ “working better” with one group over another. Although we made minor changes in the materials and processes between the two forums, the changes that we made could not have accounted for the specific differences between the groups. The changes related to clarifying some details to make the process smoother; the differences reflected distinct ways of thinking about the issues. This finding underscores the importance of gathering a diverse, regionally varied sample of Maryland citizens for the planned statewide public engagement forums. Moreover, the finding suggests that gathering informed views from demographically different constituents affords an opportunity to familiarize state authorities with locally resonant issues. By being conversant in the concerns of greatest importance to diverse groups, health authorities and elected officials will be better prepared both to shape and to communicate the content of Maryland’s future healthcare allocation framework in meaningful ways.
Our study has the advantage of the recruitment of a diverse group of participants for the qualitative pilot phase of the work. We do recognize, however, that selection bias may challenge larger-scale implementation of deliberative democratic methods. Indeed, as we have mentioned, our pilot samples were noted to be older and more likely to be female than census-derived indicators suggest are reflective of the general population in the areas from which we recruited our participants. Although we believe the census data may have been skewed for East Baltimore by inclusion of inmates in a local detention facility, who are likely to be younger and male, we cannot similarly account for differences between our Howard County sample and census data.
The deliberative methodology that we used has the advantage of allowing time for participants to offer considered responses to challenging questions. Nevertheless, the time required to participate (5 h) may be prohibitive for certain populations (e.g., hourly wage earners, families with young children) and may lead to more participation by retirees or other demographic groups with more flexible schedules. Ideally, recruitment would be done by random digit dialing or a similar method, and such recruitment would continue until a representative sample was obtained. Although ideal, this strategy is likely to be cost-prohibitive in many settings. To mitigate this problem, efforts could be made in future studies to focus some recruitment in locations, or among groups, with a younger demographic, such as college campuses.
An additional concern is the possibility of the influence of recruitment bias on the findings and on conclusions generated from a larger future set of forums in which deliberative methods are employed. Given the time commitment and the nature of the conversations, those who choose to participate may have more specific interests in the issue at hand that may lead to results that are inconsistent with the sentiments of the population at large.
This pilot work is expected to form the foundation for a Maryland statewide public engagement process on this critical topic. That process is expected to include at least 10 forums: 5 for general community members and 5 for healthcare professionals. The investigators will synthesize the feedback gained in the statewide process into a report that will be shared with the state government to inform the development of a common framework in Maryland for the allocation of scarce healthcare resources during disasters.
The authors thank Chrissie Juliano and Debbie Lee of RESOLVE for their leadership and support of the project team in the completion of this project. We would additionally like to thank Rev. Marshall Prentice, Dr. Janyne Althaus, and all the volunteer facilitators, notetakers, and participants in this work who so graciously shared their time to make the effort a success.
| 1 . | U.S. Department of Health and Human Services. HHS Pandemic Influenza Plan. Washington, DC: Author; 2005 [accessed 2014 May 14]. Available from: from http://www.flu.gov/planning-preparedness/federal/hhspandemicinfluenzaplan.pdf |
| 2 . | Adalja AA, Watson M, Waldhorn RE, Toner ES. A conceptual approach to improving care in pandemics and beyond: severe lung injury centers. J Crit Care 2013;28:318.e9–318.e15. |
| 3 . | Groeger JS, Guntupalli KK, Strosberg M, Halpern N, Raphaely RC, Cerra F, Kaye W. Descriptive analysis of critical care units in the United States: patient characteristics and intensive care unit utilization. Crit Care Med 1993;21:279–291. |
| 4 . | White DB, Katz MH, Luce JM, Lo B. Who should receive life support during a public health emergency? Using ethical principles to improve allocation decisions. Ann Intern Med 2009;150:132–138. |
| 5 . | Powell T, Christ KC, Birkhead GS. Allocation of ventilators in a public health disaster. Disaster Med Public Health Prep 2008;2:20–26. |
| 6 . | Christian MD, Hawryluck L, Wax RS, Cook T, Lazar NM, Herridge MS, Muller MP, Gowans DR, Fortier W, Burkle FM. Development of a triage protocol for critical care during an influenza pandemic. CMAJ 2006;175:1377–1381. |
| 7 . | Public Health Seattle & King County. Additional Pandemic Flu Resources: Public Engagement Project on Medical Service Prioritization during an Influenza Pandemic. Seattle, WA: Author; 2009 September 29 [accessed 2014 May 14]. Available from: http://www.kingcounty.gov/healthservices/health/preparedness/pandemicflu.aspx#footer-nav |
| 8 . | Public Engagement Pilot Project on Pandemic Influenza (PEPPPI). Citizen Voices on Pandemic Flu Choices: A Report of the Public Engagement Pilot Project on Pandemic Influenza. Lincoln, NE: University of Nebraska Public Policy Center; December 2005 [accessed 2014 May 14]. Available from: http://ppc.unl.edu/wp-content/uploads/2005/12/PEPPPI_FINALREPORT_DEC_2005.pdf |
| 9 . | Cavalier R, editor. Approaching deliberative democracy: theory and practice. Pittsburgh, PA: Carnegie Mellon University Press; 2011. |
| 10 . | Johns Hopkins Berman Institute of Bioethics. Too Many Patients, Too Few Resources. Baltimore, MD: Author; 2014 [accessed 2014 May 14]. Available from: http://www.bioethicsinstitute.org/scarcemed |
| 11 . | Fox CE, Morford TG, Fine A, Gibbons MC. The Johns Hopkins Urban Health Institute: a collaborative response to urban health issues. Acad Med 2004;79:1169–1174. |
| 12 . | Baltimore City Health Department. 2011 Neighborhood Health Profile: Perkins/Middle East. Baltimore, MD: Author; December 2011. [accessed 2014 May 14]. Available from: http://www.baltimorehealth.org/info/neighborhood2011/44%20Perkins.pdf |
| 13 . | Robert Wood Johnson Foundation/University of Wisconsin Population Health Institute. County Health Rankings & Roadmaps: Building a Culture of Health, County by County (2014 Rankings: Key Findings Report). [accessed 2014 May 14]. Available from: http://www.countyhealthrankings.org/sites/default/files/2014%20County%20Health%20Rankings%20Key%20Findings.pdf |
| 14 . | U.S. Department of Commerce, U.S. Census Bureau. State and County QuickFacts: Howard County, Maryland. Washington, DC: Author [accessed 2014 May 14]. Available from: http://quickfacts.census.gov/qfd/states/24/24027.html |
| 15 . | Hoffmaster CB, editor. Bioethics in social context. Philadelphia, PA: Temple University Press; 2001. |
| 16 . | Kleinman A. Moral experience and ethical reflection: can ethnography reconcile them? A quandary for “the new bioethics”: remaking the case for bioethics. Daedalus 1999;128:69–97. |
| 17 . | Brodwin P. The coproduction of moral discourse in U.S. community psychiatry. Med Anthropol Q 2008;22:127–147. |
Supported by a grant from the Johns Hopkins Preparedness and Emergency Response Research Center.
Author Contributions: E.L.D.B.: Study design, data analysis, and manuscript writing. H.G.: Study design and manuscript writing. M.S.-S.: Study design, data analysis, and manuscript writing. R.C.: Study design and manuscript writing. D.B.W.: Study design and manuscript writing. T.D.: Study design and study personnel training. P.B.T.: Study design and manuscript writing. A.J.L.: Study design and manuscript writing. A.R.: Study design, data analysis, and manuscript writing. R.F.: Study design, data analysis, and manuscript writing. E.S.T.: Study design, data analysis, and manuscript writing.
Author disclosures are available with the text of this article at www.atsjournals.org.
