Journal of General Internal Medicine

, Volume 25, Issue 7, pp 725–730

Religious Hospitals and Primary Care Physicians: Conflicts over Policies for Patient Care

Authors

    • Department of Family MedicineThe University of Chicago
    • Section of Family Planning, Department of Obstetrics & GynecologyThe University of Chicago
    • MacLean Center for Clinical Medical EthicsThe University of Chicago
  • Ryan E. Lawrence
    • Pritzker School of MedicineThe University of Chicago
  • Jason Shattuck
    • Internal Medicine/PediatricsWenatchee Valley Medical Center
  • Farr A. Curlin
    • MacLean Center for Clinical Medical EthicsThe University of Chicago
    • Section of General Internal Medicine, Department of Internal MedicineThe University of Chicago
Original Research

DOI: 10.1007/s11606-010-1329-6

Cite this article as:
Stulberg, D.B., Lawrence, R.E., Shattuck, J. et al. J GEN INTERN MED (2010) 25: 725. doi:10.1007/s11606-010-1329-6

ABSTRACT

BACKGROUND

Religiously affiliated hospitals provide nearly 20% of US beds, and many prohibit certain end-of-life and reproductive health treatments. Little is known about physician experiences in religious institutions.

OBJECTIVE

Assess primary care physicians’ experiences and beliefs regarding conflict with religious hospital policies for patient care.

DESIGN

Cross-sectional survey.

PARTICIPANTS

General internists, family physicians, and general practitioners from the AMA Masterfile.

MAIN MEASURES

In a questionnaire mailed in 2007, we asked physicians whether they had worked in a religiously affiliated hospital or practice, whether they had experienced conflict with the institution over religiously based patient care policies and how they believed physicians should respond to such conflicts. We used chi-square and multivariate logistic regression to examine associations between physicians’ demographic and religious characteristics and their responses.

KEY RESULTS

Of 879 eligible physicians, 446 (51%) responded. In analyses adjusting for survey design, 43% had worked in a religiously affiliated institution. Among these, 19% had experienced conflict over religiously based policies. Most physicians (86%) believed when clinical judgment conflicts with religious hospital policy, physicians should refer patients to another institution. Compared with physicians ages 26–29 years, older physicians were less likely to have experienced conflict with religiously based policies [odds ratio (95% confidence interval) compared with 30–34 years: 0.02 (0.00–0.11); 35–46 years: 0.07 (0.01–0.72); 47–60 years: 0.02 (0.00–0.10)]. Compared with those who never attend religious services, those who do attend were less likely to have experienced conflict [attend once a month or less: odds ratio 0.06 (0.01–0.29); attend twice a month or more: 0.22 (0.05–0.98)]. Respondents with no religious affiliation were more likely than others to believe doctors should disregard religiously based policies that conflict with clinical judgment (13% vs. 3%; p = 0.005).

Conclusions

Hospitals and policy-makers may need to balance the competing claims of physician autonomy and religiously based institutional policies.

KEY WORDS

religionconscienceprimary carehospital policyhealth policy

INTRODUCTION

Religious hospitals represent approximately 13% of all US community-based hospitals and provide nearly 20% of hospital beds. These facilities receive public funding in the form of Medicare and Medicaid payments and tax-exempt government bonds.1 Some religious hospitals prohibit physicians who work in their facilities from providing medical interventions that conflict with religious teaching, such as contraception, abortion, and certain end-of-life treatment options.2 General internists and family physicians provide a significant share of reproductive3,4 and geriatric care,57 so their beliefs and experiences with religious hospitals may affect many patients seeking these services.

Previous research has examined physician-patient moral conflicts that arise when patients request procedures from physicians who hold moral or religious objections to the requested procedure.8,9 But as health care is increasingly provided within institutions that constrain professional behavior,10,11 important questions emerge about physician-institution moral conflicts. When hospital policies limit physician practice based on moral or religious teachings, how do physicians respond or believe they should respond? This problem is described only in case presentations and opinion pieces in the medical literature.11,12 To our knowledge, no studies have systematically evaluated physician-institution moral conflicts over patient care.

In December 2008, the United States Department of Health and Human Services issued regulations to strengthen protections for health care providers’ rights of conscience.13 These regulations aimed to protect both religious institutions and individual providers, but they did not address the potential for conflict between the two parties. Spurring further debate, the Obama administration has moved to rescind these regulations, citing concerns about patients’ access to family planning and end-of-life services.14 In recent health reform debates, proposed policies to guarantee hospitals’ and individual providers’ rights to opt out of providing certain services on moral or religious grounds have again raised controversy.15 Despite the large number of religiously affiliated hospitals, the importance of potential physician-hospital conflict to current policy debates, and widespread attention to these issues in the lay press,16,17 little is known about the experiences of physicians working in religiously affiliated health care institutions.

We conducted this study to identify the proportion of primary care physicians who have worked in religiously affiliated hospitals or practices, their rate of conflict over religiously based policies for patient care, and their responses to such conflicts. We examined physician characteristics that have previously been associated with physician response to patient-doctor moral conflict to assess whether these would also predict physicians’ experiences and responses to conflict with religious hospitals. In a previous study, Curlin and colleagues asked a representative sample of US physicians how a physician should respond when their patient requests a procedure to which the physician has a religious or moral objection; they found that male physicians, those with higher intrinsic religiosity, and those who attend religious services more frequently are more likely to believe the physician may describe their moral objections and less likely to believe physicians are obligated to disclose all legal options or refer patients for the procedure.9 We therefore hypothesized that male physicians, those with higher intrinsic religious motivation, and those with greater participation in religious activities would similarly be less likely to experience conflict with religious health care institutions and more likely to believe physicians should comply with religiously based policies.

METHODS

In 2007 we mailed a confidential, self-administered questionnaire to a stratified random sample of 1,000 general internal medicine, family medicine, or general practice physicians in active practice, age 60 and under, drawn from the American Medical Association Physician Masterfile—a database intended to include all physicians in the US. In order to capture diverse perspectives and adequately represent Muslim, Hindu, and Buddhist physicians (who would otherwise have been too few in number to allow for robust population estimates), we used a technique developed by Lauderdale and Kestenbaum to identify people with South Asian or Arabic surnames.18,19 We divided the Masterfile into three strata: physicians with South Asian surnames, those with Arabic surnames, and everyone else. We randomly selected 250 members from each of the South Asian and Arabic surname groups and 500 from the other group. Physicians received up to three separate mailings of the questionnaire. The first included a $5 retail gift card and the third offered $30 for participation. This study was approved by the University of Chicago Institutional Review Board.

Questionnaire

This study was part of a larger survey designed to evaluate associations between physicians’ religious characteristics and their clinical practices (questionnaire available online as Appendix). The primary outcome variables for this study were whether physicians had experienced conflict with a hospital or practice over religiously based policies for patient care and physicians’ judgments about how such conflicts should be handled. Measures were developed through an iterative process based on the input of expert colleagues and clinicians from a diverse range of medical specialties and religious orientations.

Physicians were asked, “Have you ever taken care of patients in a religiously affiliated hospital or practice?” (Yes/No) Those who answered “Yes” were asked to indicate the religious affiliation of that hospital/practice (Jewish, Roman Catholic, Christian non-Catholic, other) and whether they have “ever had a conflict with that practice/hospital regarding its religiously based policies for patient care?” (Yes/No)

In addition, all physicians were asked, “What should a physician do if he/she believes that a patient needs a medical intervention, and the hospital in which the physician works prohibits that intervention because of its religious affiliation?” Response options were: (1) provide the intervention openly, even if doing so risks the physician’s job or hospital privileges, (2) provide the intervention discretely in order to avoid risking the physician’s job or hospital privileges, (3) encourage the patient to seek the intervention at another hospital, and (4) recommend another treatment option that is permitted at this hospital. Responses to this item were dichotomized to distinguish those who believe physicians should comply with religiously based hospital policies by referring the patient elsewhere or recommending an alternate treatment and those who believe physicians should provide an intervention despite its prohibition by hospital policy.

Primary predictor variables were physician sex and measures of physicians’ religious characteristics. Religious affiliation was categorized as no religion, Hindu, Muslim, Catholic/Orthodox, evangelical Protestant, non-evangelical Protestant, and other religion. Jewish (n = 16) and Buddhist (n = 5) physicians were included among “other” because of the small number of respondents. Intrinsic religious motivation—the extent to which an individual embraces their religion as the master motive that guides and gives meaning to their life20—was measured using seven items (see Appendix, item 7, subparts a, b, c, d, e, f, and h) derived from the Hoge Intrinsic Religious Motivation Scale21. Responses, which ranged from 1 (strongly agree) to 4 (strongly disagree), were averaged to create a scale with high internal reliability (Cronbach alpha = 0.94) from which we created an ordered three-category (high, medium, low) variable. Participatory religiosity was measured as frequency of attendance at religious services and was categorized as never, once a month or less, or twice a month or more. Finally, physicians were asked to what extent they consider themselves a spiritual person and to what extent a religious person. Reponses, ranging from 1 (very) to 4 (not at all), were dichotomized, and respondents were classified as religious, spiritual not religious, or neither. These variables and response groupings were derived from prior research designed to develop valid measures of physician religious characteristics.9,22

Additional predictors were physicians’ age, US Census geographic region (Northeast, Midwest, South, West, or Puerto Rico), specialty (Internal Medicine, Family Medicine, or General Practice), and immigration history (born in the US, immigrated to the US as a child, or immigrated to the US as an adult).

Statistical Analysis

In order to make estimates for the national population of primary care physicians, we adjusted statistically for the stratified sample design and different non-response rates across groups.23 We first calculated a sample weight that accounted for the unequal selection probabilities within the three sample strata (South Asian surname, Arabic surname, and everyone else); we then calculated a weight taking into account non-response rates by physician sex, geographic region, and specialty. A final weight was calculated by multiplying the stratification and non-response weights together. This was used to generate adjusted population estimates for each of the outcome measures.

We created a predictor variable to test for the relationship between physician-institution religious congruence and reported experience of conflict. Physicians who reported working in a religious hospital or practice of the same religious denomination as their own personal religious affiliation were considered congruent; those who worked in a religious hospital or practice of one denomination and identified with any other religious affiliation (or none) were considered incongruent.

We used the chi-square test and multivariate logistic regression to examine associations between each predictor and each outcome measure. All analyses were conducted with Stata SE v10 statistical software (College Station, TX).

RESULTS

Approximately 12% of the questionnaires were returned undeliverable. From the remaining sample, the response rate was 51% (446/879). Response rates varied among the three samples: 55% (246 respondents/450 eligible) of the general sample responded, 49% (104/212) of those with South Asian surnames responded, and 44% (96/217) of those with Arabic surnames responded. There was no significant variation in response rate by gender, geographic region, or specialty. Respondent characteristics are shown in Table 1. Counts vary slightly due to item non-response; missing data were excluded from analysis.
Table 1

Demographics of Survey Respondents (n = 446)

Respondent characteristics

n (%)

Sex

 

 Female

176 (39)

Race

 

 Asian

191 (44)

 Black or African-American

18 (4)

 Hispanic or Latino

23 (5)

 White or Caucasian

192 (44)

 Other

13 (3)

Age

 

 26–29

107 (24)

 30–34

119 (27)

 35–46

112 (25)

 47–60

108 (24)

Immigration history

 

 Born in the US

216 (50)

 Immigrated to US as a child or adult

217 (50)

Specialty

 

 Family Medicine or General Practice

118 (26)

 Internal Medicine

328 (74)

Region

 

 South

125 (28)

 Midwest

110 (25)

 Northeast

129 (29)

 West

72 (16)

 Puerto Rico

10 (2)

Religious affiliation

 

 No religion

50 (11)

 Hindu

93 (21)

 Muslim

76 (17)

 Catholic/Orthodoxa

94 (21)

 Protestant, evangelical

26 (6)

 Protestant, non-evangelical

71 (16)

 Other religionb

35 (8)

Intrinsic religious motivation

 

 Low

153 (35)

 Medium

120 (27)

 High

170 (38)

Attendance at religious services

 

 Never

53 (12)

 Once a month or less

244 (55)

 Twice a month or more

147 (33)

Religious/Spiritual

 

 Neither

94 (21)

 Spiritual not religious

101 (23)

 Religious

248 (56)

aIncludes Roman Catholic (n = 83) and Orthodox (n = 11)

bIncludes Buddhist (n = 5), Jewish (n = 16), and other religions (n = 14)

Many respondents (n = 177) reported having taken care of patients in a religiously affiliated institution (40% of unadjusted sample; 43% adjusting for sampling design). Further percentages presented here are survey design-adjusted in order to present population estimates. When asked the religious affiliation of the hospital or practice, 115 (design-adjusted percentage, 31%) indicated Roman Catholic, 49 (10%) Christian, non-Catholic, and 18 (3%) Jewish. Among doctors who reported working in religiously affiliated institutions, approximately one in five (19%, n = 33) had experienced a conflict with the institution’s religiously based patient care policies. The vast majority (86%, n = 365) of physicians indicated that when such conflicts arise doctors should encourage patients to seek the recommended intervention at a hospital where the intervention is not prohibited. Fewer (10%, n = 49) endorsed recommending another treatment option that is permitted at the religiously affiliated institution, and only 4% (n = 23) endorsed providing the prohibited intervention (either openly or discretely) in violation of hospital policy.

Table 2 shows the likelihood of having worked in a religiously affiliated hospital or practice by physician characteristics. Women and men had similar likelihood of working in a religiously affiliated institution. Religious affiliation was not significantly associated with having worked in a religious hospital. Doctors with high intrinsic religious motivation were more likely to report having worked in a religiously affiliated institution (82/170, design-adjusted percentage 52%) compared to those with medium (41/115, 34%) or low (53/151, 40%) intrinsic religious motivation (p = 0.03). Similarly, those who reported attending religious services at least twice a month were more likely to report having worked in a religious hospital (76/147, 55%) than those attending services once a month or less (83/239, 35%) or never (17/50, 42%, p = 0.006). However these associations did not remain statistically significant when adjusted for physician specialty and demographics.
Table 2

Prevalence of Physicians’ Having Worked in a Religious Institutiona

Respondent Characteristics (no. responding to itemb)

Have worked in a religiously affiliated practice or hospital (no. responding “yes”b)

 

Bivariate

Multivariate

 

n(%)

p(χ2)

OR (95% CI)c

Sexd

   

 Female (173)

70 (45)

0.72

1.0 referent

 Male (265)

107 (43)

 

0.9 (0.5–1.4)

Aged

   

 26–29 (106)

35 (34)

0.29

1.0 referent

 30–34 (116)

46 (41)

 

1.6 (0.8–3.2)

 35–46 (110)

47 (49)

 

1.9 (0.9–4.1)

 47–60 (106)

49 (46)

 

1.3 (0.6–3.1)

Geographic regiond

   

 South (121)

46 (45)

0.37

1.0 referent

 Midwest (109)

55 (51)

 

1.2 (0.6–2.3)

 Northeast (127)

49 (41)

 

0.6 (0.3–1.3)

 West (71)

24 (36)

 

0.6 (0.3–1.2)

 Puerto Rico (10)

3 (28)

 

0.5(0.1–2.3)

Religious affiliatione

   

 No religion (49)

20 (47)

0.69

1.0 referent

 Hindu (90)

31 (44)

 

1.0 (0.4–2.5)

 Muslim (75)

24 (33)

 

0.5 (0.2–1.4)

 Catholic/Orthodox (93)

48 (49)

 

1.0 (0.5–2.3)

 Protestant, evangelical (26)

13 (48)

 

1.0 (0.3–2.8)

 Protestant, non-evangelical (70)

27 (38)

 

0.5 (0.2–1.2)

 Other religion (35)

14 (38)

 

0.6 (0.2–1.8)

Intrinsic religious motivatione

   

 Low (151)

53 (40)

0.03

1.0 referent

 Medium (115)

41 (34)

 

0.7 (0.4–1.4)

 High (170)

82 (52)

 

1.6 (0.9–2.9)

Attendance at religious servicese

   

 Never (50)

17 (42)

0.006

1.0 referent

 ≤Once a month (239)

83 (35)

 

0.8 (0.4–1.7)

 ≥Twice a month (147)

76 (55)

 

1.6 (0.7–3.7)

aPercentages reflect survey-design-adjusted estimates. Example interpretation: after adjusting for survey design, 43% of male physicians (n = 107) report having worked in a religiously affiliated hospital or practice. In multivariate analysis, men have 0.9 times (95% confidence interval = 0.5–1.4) the odds of having worked in a religious practice or hospital compared to women

b See Appendix, question 23

cOdds ratio (95% confidence interval) for odds of answering “yes”

dMultivariate model controls for specialty, demographics (age, geographic region, immigration history), and religious characteristics (religious affiliation, intrinsic religious motivation, attendance at religious services, and identification as religious/spiritual)

eMultivariate model controls for specialty and demographics

Table 3 displays physician experience of and response to conflicts with religious institutions by physician characteristics. Women were twice as likely as men to have experienced conflict with religiously based hospital policies (19/66, design-adjusted percentage 29%, vs. 14/100, 14%; p = 0.04), but this association did not remain significant after adjusting for other covariates. In multivariate analyses, compared to the youngest physicians (ages 26–29), those of older age groups were less likely to report conflict with a religious hospital or practice, though no linear trend in strength of association was apparent. Odds ratio (OR) [95% confidence intervals (95% CI)] were, for comparison with physicians, 30–34 years: 0.02 (0.00–0.11); for 35–46 years: 0.07 (0.01–0.72); for 47–60 years 0.02 (0.00–0.10)]. Compared with doctors practicing in the South, those in the Midwest and West were more likely to report conflict [OR (95% CI), 6.7 (1.2–38), and 12.1 (1.2–119), respectively].
Table 3

Physician Experiences and Beliefs Regarding Conflicts with Religiously Based Institutional Policiesa

Respondent characteristics

Have had conflict with religiously based policies

Believe doctors should provide indicated treatment despite policy

 

No. responding to itemc

No. responding “yes”c

No. responding to itemd

No. responding "openly” or “discretely”d

  

Bivariate

Multivariate

 

Bivariate

Multivariate

  

n (%)

p(χ 2)

OR (95% CI)

 

n(%)

p(χ 2)

OR (95% CI)

Sexe

        

 Female

66

19(29)

0.04

1.0 referent

174

7(3)

0.57

1.0 referent

 Male

100

14(14)

 

0.5(0.1–2.1)

263

16(5)

 

1.4(0.4–5.3)

Agee

        

 26–29

35

13 (46)

0.005

1.0 referent

107

6 (4)

0.35

1.0 referent

 30–34

43

6 (12)

 

0.02(0.00–0.11)g

118

8 (7)

 

2.4(0.4–15.5)

 35–46

43

9 (20)

 

0.07(0.01–0.72)g

108

3 (2)

 

0.7(.005–0.9)g

 47–60

46

5 (10)

 

0.02(0.00–0.10)g

104

6 (5)

 

0.2(0.03–1.9)

Geographic regione

        

 South

43

6 (13)

0.56

1.0 referent

121

6 (4)

0.69

1.0 referent

 Midwest

52

11 (25)

 

6.7 (1.2–37.7)g

109

3 (2)

 

0.3(0.05–1.4)

 Northeast

46

10 (19)

 

3.5(0.7–18.6)

127

7 (5)

 

0.6(0.1–2.4)

 West

23

6 (26)

 

12.1 (1.2–119.0)g

70

6 (6)

 

1.1(0.2–6.3)

 Puerto Rico

3

0 (0)

 

n/a

10

1 (9)

 

5.8(0.7–51)

Religious affiliationf

        

 No religion

19

7 (33)

0.13

1.0 referent

50

6 (13)

0.009

1.0 referent

 Hindu

30

8 (27)

 

0.8 (0.2–3.6)

92

5 (3)

 

0.1 (0.0–0.6)g

 Muslim

23

6 (39)

 

2.4 (0.4–15.2)

74

3 (3)

 

0.2 (0.0–1.0)g

 Catholic/Orthodox

44

7 (18)

 

0.4 (0.1–1.9)

90

2 (3)

 

0.1 (0.0–0.8)g

 Protestant, evangelical

12

1 (7)

 

0.1 (0.0–1.1)

26

2 (4)

 

0.2 (0.0–4.0)

 Protestant, non-evangelical

25

3 (14)

 

0.4 (0.1–2.2)

70

1 (<1)

 

0.0 (0.0–0.2)g

 Other religion

14

1 (2)

 

0.0 (0.0–0.3)g

35

4 (10)

 

0.7 (0.1–4.3)

Intrinsic religious motivationf

        

 Low

50

13 (26)

0.10

1.0 referent

153

10 (6)

0.44

1.0 referent

 Medium

39

12 (25)

 

1.2 (0.3–4.0)

115

4 (4)

 

0.5 (0.1–1.9)

 High

77

7 (11)

 

0.4 (0.1–1.2)

167

9 (3)

 

0.5 (0.1–2.2)

Attendance at religious servicesf

        

 Never

16

6 (38)

0.08

1.0 referent

52

7 (13)

0.02

1.0 referent

 ≤Once a month

78

12 (13)

 

0.06 (0.01–0.29)g

237

6 (2)

 

0.1 (0.0–0.7)g

 ≥Twice a month

72

14 (19)

 

0.22 (0.05–0.98)g

146

10(5)

 

0.4 (0.1–2.0)

aPercentages reflect survey-design-adjusted estimates. Denominator for column 2 ("Have had conflict") is respondents who have worked in a religious hospital (n = 177), while for column 3 ("Believe doctors should provide") it is all respondents (n = 446). Example interpretation: after adjusting for survey design, 14% of male physicians (n = 14) report having experienced conflict with religiously based institutional policies, and 5% (n = 16) believe physicians should provide an indicated treatment despite hospital policy prohibiting it. In multivariate analysis, men have 0.5 times (95% confidence interval = 0.1–2.1) the odds of having experienced conflict compared to women.

cSee Appendix, question 23b

dSee Appendix, question 22

eMultivariate model controls for specialty, demographics (sex, age, geographic region, immigration history), and religious characteristics (religious affiliation, intrinsic religious motivation, attendance at religious services, and identification as religious/spiritual)

fMultivariate model controls for specialty and demographics

gp < 0.05

Neither religious affiliation nor physician-institution congruence was significantly associated with having experienced conflict with religiously affiliated institutions. Conflict rates were roughly the same among physicians who shared the religious denomination of their hospital or practice (7/37, design-adjusted percentage 21%) and those who did not (26/130, 18%, p = 0.73). However, doctors with no religion were roughly three times as likely as those with religious affiliations to believe that physicians should provide an intervention despite its prohibition by hospital policy (6/50, 13%, vs. 17/387, 3%; p = 0.005). In adjusted analyses, compared with physicians who never attend religious services, those who attend up to once per month and those who attend twice a month or more were less likely to report conflict with religiously affiliated institutions, though no trend in strength of association is seen [OR (95% CI) for comparison with attend up to once a month: 0.06 (0.01–0.29); for attend twice a month or more 0.22 (0.05–0.98)]. Respondents who attend religious services up to once per month were less likely to believe that doctors should provide an intervention prohibited by hospital policy compared to those who never attend [OR (95% CI) 0.1 (0.0–0.7)]. Intrinsic religious motivation was not significantly associated with experience of or responses to conflict.

DISCUSSION

This study empirically examined physician-hospital conflicts over religiously based patient care policies. We found that almost half of primary care physicians have worked in a religiously affiliated hospital or practice, and among these physicians, approximately one in five has had conflicts with the institution’s religiously based patient care policies. Most primary care physicians believe that when a physician’s clinical judgment conflicts with religiously based policies, the physician should refer the patient to another hospital. Controlling for potential confounders, we found no association between religious characteristics and whether or not a physician had worked in a religious hospital. However, among physicians who have worked in religious hospitals, younger physicians and those who never attend religious services are somewhat more likely to report conflict with religiously based policies.

As a result of widespread hospital consolidations many patients, and perhaps particularly those in underserved communities, have fewer choices regarding where to receive health care.24 Policy debates about religious hospitals have largely focused on the tension between patients’ rights to access services and hospitals’ rights to practice within their moral tradition. Our study does not measure the effect of physician-hospital conflicts on patients’ access to services. It does, however, suggest that patients cared for in a religious hospital or practice who seek time-sensitive but restricted interventions—such as emergency contraception—may face delays as their physicians transfer or refer them to non-religious institutions.25 Whether these delays are seen as harmful to the patient depends on one’s beliefs about the intervention itself; even among the authors of this paper, judgments vary.

Our findings also draw attention to a lacuna in the recent federal regulations written to protect individual providers and health care institutions from discrimination based on their religious beliefs. Namely, such policies do not address the relatively common problem of conflicts between physicians and religious institutions.

This study’s strengths include a nationally representative sample of primary care physicians and adequate inclusion of minority religions. The religious characteristics and moral beliefs of our sample are similar to those reported in previous surveys of US physicians.9 However, the study has several important limitations. Questionnaire items on physician-hospital conflict were not formally pilot-tested prior to the study, so their reliability is not established. Survey questions and responses used to characterize physician religiosity reflect those used in related research.9,22,26,27 The failure to see directional trends where they might be predicted (for example, we might have expected more frequent attendance at religious services to be associated with a lower odds of conflict with religious hospital policies than infrequent attendance) may be due to chance given the small sample size for sub-group analysis. However, this failure may call into question the usefulness of these results. Limited survey space prevented us from seeking richer, qualitative descriptions of the conflicts physicians experienced, so physician motivation, patients’ roles, and clinical details of the conflicts remain open areas for future research. Our response rate was 51%, and we lack information about non-respondents, who may differ from respondents in ways that could bias the findings. African-American and Latino physician numbers were small. Furthermore, while this study describes primary care physician experiences working at religiously affiliated hospitals and practices, it does not compare these to experiences working in non-religiously affiliated sites. Finally, the opinions and experiences of specialist physicians, hospital staff, and patients are not represented in this study. Future work will examine the experiences and beliefs of obstetrician/gynecologists, the specialty perhaps most likely to encounter religious prohibitions on procedures within its scope of practice.

Notwithstanding these limitations, these results suggest that a significant minority of primary care physicians working in religiously affiliated health care institutions has faced conflict over religious policies for patient care. Hospital administrators may wish to better involve physicians in the policy-making process, communicate policies more clearly, and develop means of hearing and accommodating physician concerns in order to reduce conflict and its impact on patient care.28 Physicians should also consider informing patients about religiously based institutional policies if an admission is not urgent and other hospital choices are available. Policy-makers may find physicians’ experiences reported here useful in addressing the role of religious institutions in the delivery of health care.

Acknowledgments

Funding for this work was provided by the Greenwall Foundation, New York, NY, and the National Center for Complementary and Alternative Medicine (1 K23 AT002749). Mr. Lawrence’s time was supported by the University of Chicago Pritzker School of Medicine Summer Research Program. The study was presented at the annual meeting of the North American Primary Care Group, November 18, 2008.

We thank Joshua Kellemen for assistance with data collection and management; Kenneth Rasinski for assistance with data analysis; and Bernard Ewigman for manuscript review.

Conflicts of Interest

None disclosed.

Supplementary material

11606_2010_1329_MOESM1_ESM.pdf (410 kb)
ESM 1(PDF 410 kb)

Copyright information

© Society of General Internal Medicine 2010