Patient Views on Religious Institutional Health Care

IMPORTANCE Federal protections allow health facilities to limit options to patients on the basis of religious values. Little is known about whether US adults consider religious affiliation when selecting facilities and whether they agree with such limitations. OBJECTIVE To understand patient views on religious institutional care. DESIGN, SETTING, AND PARTICIPANTS This is a population-based, cross-sectional survey study of US adults recruited from the probability-based AmeriSpeak Omnibus panel available from NORC (formerly the National Opinion Research Center) at the University of Chicago. Surveys were administeredviainternetortelephoneduringa3-dayfieldingperiodinNovember2017.Dataanalysis was performed from January 2018 to October 2019. EXPOSURES Participant characteristics, including religiosity measures. gender existed, performed, applied


Introduction
The nature of US health care is shifting, in part because of the growing religious ownership sector. As of 2016, 18.5% of hospitals were religiously affiliated: 9.4% were Catholic-owned nonprofit hospitals, 5.1% were Catholic-affiliated hospitals, and 4.0% were other religious nonprofit hospitals. 1 Catholic hospitals in particular have demonstrated significant growth recently; between 2001 and 2016, the number of acute care hospitals that were Catholic owned or affiliated grew by 22%, while the overall number of acute care hospitals decreased by 6% and the number of other nonprofit religious hospitals decreased by 38.3%. 1 In 2016, 10 of the top 25 health care systems were Catholic sponsored. 1 Almost half (46%) of all US Catholic hospitals are located in the Midwest, 2

and 46
Catholic hospitals are designated as sole community hospitals because of their remote location from other major medical centers. 1 Attendance at religious health care facilities can affect a patient's access to services because of religious interpretations about care designated by the institution. Specific to Catholic health care facilities, clinicians are expected to abide by the Ethical and Religious Directives for Catholic Health Care Services, 3 which places limitations on reproductive and end-of-life-care methods on the basis of the church's moral teachings. Prior evidence 4 has highlighted restrictions to care in Catholic facilities specific to contraception, sterilization, miscarriage management, and abortion. Recent media reports have highlighted conflicts in care with respect to transgender health and medical aid in dying. In other non-Catholic religious hospitals, contraceptive and sterilization services are generally provided, whereas abortion care is often restricted. 5 In 1973, the US federal government provided the first protection to health care entities, known as the Church Amendment, 6 to allow institutions to refuse to provide services that conflict with their religious beliefs or moral values. Such protections allow religious refusals of care to be implemented at the institutional level, without formally establishing religious doctrine as the basis of health care.
Since that time, other conscience protections have emerged. In January 2018, the Trump administration founded the Conscience and Religious Freedom Division in the US Department of Health and Human Service Office for Civil Rights, to "protect the fundamental and unalienable rights of conscience and religious freedom." 7 The Office for Civil Rights issued a final conscience rule 7 in May 2019 that broadens and enforces the right of religious health care entities-hospitals, clinics, insurance companies, and others-to invoke their institutional conscience to restrict options. In November 2019, a federal judge voided this rule, writing that the "stated justification for undertaking rule making in the first place-a purported 'significant increase' in civilian complaints relating to the conscience provisions-was factually untrue." 8 It is unknown to what extent patients consider religious affiliation when selecting their health care facilities and whether they believe institutions have the right of conscience over their own medical desires and needs. Prior surveys [9][10][11][12][13]

Methods
We created a national cross-sectional survey of US adults that was administered by NORC (formerly the National Opinion Research Center) at the University of Chicago in November 2017. We used the coverage of difficult-to-survey population segments, such as rural and low-income households. 15 The omnibus service ensures responses from at least 1000 adults aged 18 years and older after fielding the survey during a 3-day period (Friday through Sunday). 15 Panel members receive a survey request that is not specific to the survey topic and are given the option to complete online or by telephone. 15 All data were deidentified, and the Colorado Multiple Institutional Review Board deemed this study exempt with waiver of consent granted. This study follows the American Association for Public Opinion Research (AAPOR) reporting guideline. 16 We focused survey development on understanding patient views of religious institutional health care. Our survey was informed by prior surveys, [9][10][11][12][13] and questions were added or modified to extend queries related to religious institutional health care affiliations. We used expert panel review and piloted the preliminary draft with 5 lay individuals to ensure readability and absence of ambiguity. We first asked participants to select any and all considerations when selecting a health care facility and subsequently asked them to rate the most important consideration. Next, we specifically asked their preference about whether the health care facility they attend has a religious affiliation with the option to respond, "I do not care whether the hospital or clinic is or is not religiously affiliated," "I prefer that the hospital or clinic is religiously affiliated," or "I prefer that the health care institution is not religiously affiliated." If they selected either of the latter 2 options, we used branching logic and asked "why?" with a free-text response option. Because we were interested in understanding how patients consider the implications of institutional conscience, we asked the We used SPSS statistical software version 25 (IBM) to calculate descriptive frequencies and compare gender responses using 2-sided χ 2 analyses. If significant gender differences existed (with significance set at P Յ .05), we sought to investigate independent factors associated with the survey query. To do so, we calculated separate logistic regression models for male and female participants including all relevant variables and reported adjusted odds ratios (aORs) and 95% CIs. We calculated survey weights according to gender and region using US Census data 17 and applied them to our analyses. We also compared characteristics of responders with those of nonresponders. For openended queries, we used inductive thematic coding to categorize primary responses. The first coder created categories based on emerging themes and applied them to all responses. A second coder received the list of inductive codes and independently applied them to all the responses. The 2 coders then met to discuss any disagreements on categorizations, and final codes were assigned.
Responses were similar across genders.
When we specifically asked whether they prefer religious affiliation when selecting a health facility, most participants (71.3%) reported they do not care whether the facility is religiously affiliated, 13.4% prefer a religious affiliation, and 15.3% prefer no religious affiliation ( were less likely to agree with this statement (Table 3). In comparison, no factors emerged as significant among women. Applying survey weights had no effect on our models.
Among the 1024 respondents who believe that their health choices should take priority over an institution's religious affiliation, 865 provided comments. Thematic responses included reference to personal choice and/or autonomy over one's own body (60.6%), that an institution's role should be focused on health over religion (24.0%), that an institution's religious affiliation may not be aligned  that health concerns will be prioritized over religion (4.2%), and personal preference (1.1%).  The only statistically significant difference between women and men was for proximity of the facility (P = .01). This discordance between how most patients choose health facilities and their beliefs about how they should receive care suggests a general lack of understanding specific to the notion of institutional conscience and may serve as the basis for conflicts in care.

JAMA Network Open | Public Health
Being a woman in and of itself was associated with support for personal autonomy over institutional conscience. This likely reflects that religious restrictions to care are of greater concern for women because of reproductive care restrictions and explains why so many commented about concerns for personal choice and/or autonomy over one's own body. 4 We were surprised that there were few other factors associated with the support for personal autonomy over institutional conscience among men. Not surprisingly, we found that men who do not associate with a religion were more likely to value personal autonomy over institutional conscience. In contrast, those who reported frequent attendance at a religious facility and/or a higher number of household members were less likely to share those values. The finding that men in the Midwest were more likely than those in the West to value their autonomy may reflect the higher proportion of Catholic Strengths of our study include that we specifically targeted concerns for religious health care, analyzed religiosity measures, and provided qualitative themes. Importantly, AmeriSpeak uses probability-based panels, which are becoming the standard given that traditional household surveys are less feasible and tend to differentially exclude many people. 15

Limitations
Our study also has limitations. Although we had a diverse sample, we still oversampled white individuals with higher education and higher income, and so our findings may be less representative of the views and opinions of underrepresented groups, including black and Hispanic populations.
Our cumulative survey response rate was also low (7.3%), which is accounted for, in part, by the way in which AmeriSpeak obtains their panels and response rates. 16 Unlike other surveys that often remove panel members if they do not respond to surveys when calculating response rates, AmeriSpeak includes panel members who do not respond. In addition, the omnibus survey service has a short fielding period of only 3 days. Although there may be concerns for responder bias based on the subject of our survey, participants were not informed of this before survey initiation.

Conclusions
Our findings demonstrate that most patients place great emphasis on their autonomy, effectively disagreeing with ongoing protections for institutions to restrict care on the basis of their religious or moral values. Advocacy efforts are needed to enact legislation that counterbalances protections for institutions with protections for patients. Because women are disproportionately affected by religious restrictions to care, as are LGBTQIA (lesbian, gay, bisexual, transgender, queer, intersex, and asexual) patients and those in rural settings, 24 advocates must work toward antidiscriminatory policies and legislation. In Washington state, legislation has passed that enforces all hospitals to report restrictions to care on their websites. 25 Because some patients in religious settings may not have other reasonable or viable options for health care access and/or may be faced with lifethreatening conditions or need medically indicated care, stronger emergency care protections are urgently needed. Broader consideration should also be given for protections that ensure provision of medically indicated care, even in nonemergent settings.