Bivariate analysis of agreement by medical specialty: A, A physician should ask about a patient's religious and spiritual beliefs. B, A physician should say a silent prayer for a patient.
Monroe MH, Bynum D, Susi B, Phifer N, Schultz L, Franco M, MacLean CD, Cykert S, Garrett J. Primary Care Physician Preferences Regarding Spiritual Behavior in Medical Practice. Arch Intern Med. 2003;163(22):2751-2756. doi:10.1001/archinte.163.22.2751
Copyright 2003 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2003
Knowledge of physician attitudes and preferences regarding religion and spirituality in the medical encounter is limited by the nonspecific questions asked in previous studies and by the omission of specialties other than family practice. This study was designed to determine the willingness of internists and family physicians to be involved with varying degrees of spiritual behaviors in varied clinical settings.
The study was a multicenter, cross-sectional, nonrandomized design recruiting physicians from 6 teaching hospitals with sites in North Carolina, Vermont, and Florida. A self-administered survey was used to explore physicians' willingness to address religion and spirituality in the medical encounter. Data were gathered on the physicians' religiosity and spirituality and sociodemographic characteristics.
Four hundred seventy-six physicians responded, for a response rate of 62.0%. While 84.5% of physicians thought they should be aware of patients' spirituality, most would not ask about spiritual issues unless a patient were dying. Fewer than one third of physicians would pray with patients even if they were dying. This number increased to 77.1% if a patient requested physician prayer. Family practitioners were more likely to take a spiritual history than general internists.
Most primary care physicians surveyed would not initiate any involvement with patients' spirituality in the medical encounter except for the clinical setting of dying. If a patient requests involvement, however, most physicians express a willingness to comply, even if the request involves prayer.
THE UNITED STATES is a religious and spiritually oriented country.1 Within this setting, the development of an increasingly holistic approach to health and medical care is challenging the separation of religion and modern medicine. In the medical literature, there is considerable interest and debate on how patients' religion and spirituality should be addressed.2- 7 From the patient perspective, the evidence to date suggests that many adults want physicians to be involved at some level in their spirituality, ranging from just being aware of their beliefs to actually praying with them.8- 14 In response to these data and the growing but controversial body of studies15,16 on religion's association with improved health outcomes, some have advocated the incorporation of a spiritual needs assessment into routine patient care.4,15,17- 19
Most of what we know about physician attitudes and preferences regarding religion and spirituality in the medical encounter comes from surveys of family practitioners. As with patient surveys, these studies14,20- 22 show that most physicians believe they should be aware of patients' religious or spiritual needs and sometimes address them. Given the limited scope of the survey questions previously asked, no data exist on whether attitudes are dependent on the clinical setting. Also, given many patients' expressed desire for prayer by their physician, it is not clear to what extent physicians are willing to fulfill this spiritual need. Finally, what we know about physicians' beliefs is limited to family practitioners, with little data available that describe other specialties.
As part of the Religion and Spirituality in the Medical Encounter study, we sought to explore physicians' attitudes and preferences regarding their response to patients' religiosity and spirituality in various clinical settings. Furthermore, we sought to include general internists, as they have been omitted in past reports.
The Religion and Spirituality in the Medical Encounter study is a multicenter, cross-sectional observational study using a structured survey of physician attitudes and preferences regarding religion and spirituality in the medical encounter. A companion patient survey was conducted and is reported elsewhere.23 The institutional review boards of each participating institution approved the study design. Physician participants were recruited from 6 sites, including Carolinas Medical Center, Wake Medical Center, Moses Cone Hospital, Brody School of Medicine at East Carolina University, University of Vermont College of Medicine, and University of Florida College of Medicine, Gainesville. Physicians surveyed included year 2000 internal medicine and family practice faculty and residents at participating programs, as well as those who completed their internal medicine residency during the preceding 5 years.
Surveys were mailed or hand-delivered to the eligible physicians. If the surveys were not returned in 3 weeks, a second mailing was sent. The participants were given another 3 weeks in which to respond before an attempt was made to contact them by telephone to remind them of the survey.
The physician survey instrument was developed by the Religion and Spirituality in the Medical Encounter study group and consists of 90 items, including demographics, a spiritual assessment, past history of and desire for education on spirituality in medicine, and preferences regarding degrees of involvement in patients' religious and spiritual life.
Demographics included age, sex, race, marital status, religious affiliation, primary specialty, and current level of training or practice setting. We measured religiosity and spirituality by incorporating the Spiritual Well-Being scale (SWB) and by quantifying worship attendance. The SWB has been previously validated and used in different research contexts and populations.24 It is 20 items long, uses a Likert scale, and assesses a person's spiritual well-being as he or she perceives it in a religious sense and an existential sense. Each item is scored from 1 to 6, with total scores ranging from 20 to 120. Higher scores represent greater spiritual well-being.
We developed a series of questions regarding physician attitudes and preferences toward involvement in their patients' religious and spiritual life based on a literature review and several focus-group discussions. The responses are based on a 5-point Likert scale and address physicians' perceived need to be aware of or play a role in their patients' religious and spiritual life. More specific attitudes were assessed by varying the clinical setting (a routine office visit, a hospitalized patient, and a patient near death) and the level of physician involvement (asking the patient about his or her religious and spiritual beliefs, saying a silent prayer for the patient, and praying with the patient). The same graded questions were repeated with the preface "If the patient requests."
We used statistical analysis to describe demographic characteristics, SWB scores, worship attendance, prior education and training, and physician attitudes. Bivariate analyses, using χ2 or t tests where appropriate, were performed to identify any associations between physician characteristics and physician preferences for involvement in patients' religious and spiritual life. Logistic regression analysis was performed. The outcome variables of interest were physician preferences concerning religious and spiritual behavior in specific medical settings. Likert scales were dichotomized so that strong agreement and agreement equaled "agree" and the remaining answers ("neutral," "disagree," and "strongly disagree") were represented by "other." Each analysis was performed controlling for SWB score, specialty, age, sex, marital status, and spiritual training during residency. The SWB score was analyzed as a trichotomous variable for which the lowest score grouping was set up as a dummy variable. Years in practice and resident status correlated highly (R>0.6) with age and were not included in the regression model. All statistical analyses were performed with Stata 6.0 (Stata Corp, College Station, Tex).
Of 768 physicians surveyed, 476 responded, for a response rate of 62.0%. Demographic characteristics are shown in Table 1. About three quarters of physicians surveyed were white, and slightly more than one third were female. Fifty-eight percent were general internists. Almost half of the physicians were Protestant, and a similar percentage were residents or fellows.
Physician preferences for participating in spiritual behaviors according to clinical setting are shown in Table 2. While 84.5% of physicians agreed that they should be aware of a patient's religious and spiritual beliefs, fewer than one third agreed that they should ask about them during a routine office visit. This percentage increased with more acute clinical settings. During a routine office visit, only 14.2% agreed that they should say a silent prayer for a patient and 5.9% agreed that they should pray with a patient. Again, as the clinical setting worsened, more physicians were willing to participate in prayer behaviors, although still fewer than one third were willing to pray with a patient even if the patient were dying.
Respondents were much more likely to agree to participate in prayer if the patient specifically requested such behavior. For example, in the setting of a routine office visit, 55.6% of physicians were agreeable to praying with a patient at the patient's request, compared with only 5.9% willing to do so without the request. Similar results were seen for the behavior of silent prayer and for patients who were hospitalized or near death. Despite being presented with the circumstance of a patient requesting prayer in the situation of near death, 22.9% of physicians still disagreed with prayer participation.
In bivariate analysis, family practitioners consistently wanted to be aware of patients' beliefs more than internists in all settings and were more likely to agree to silent prayer for hospitalized and dying patients (Figure 1). High scores on the SWB and higher attendance of organized worship services were consistently associated with physicians being more agreeable to asking patients about their beliefs, as well as the desire to pray for or with patients in every setting. The characteristics associated with physician agreement to patient requests were a higher score on the SWB (P≤.001 in every setting) and greater organized worship attendance (P≤.001). Family practitioners were more likely to agree to patient requests for silent prayer (P≤.03) and active prayer (P≤.001) during an office visit, silent prayer (P≤.008) and active prayer (P≤.001) in the hospital, and silent prayer (P≤.009) and active prayer (P≤.003) in the near-death setting. Family practice physicians were also more likely to have had training in addressing spiritual issues with patients during residency (31.4% compared with 15.4% of internists, P≤.001). There were no associations between years in practice, sex, race, and marital status with physicians' agreement to inquire about patients' beliefs or participate in prayer.
Logistic regression analysis results are summarized in Table 3. Internists were less likely to agree with asking about patients' religious and spiritual beliefs in all clinical settings. High SWB scores were associated with agreement to ask about patients' beliefs and were the only consistently significant association with physician preference to pray with patients. Higher SWB score was also the most significant association with physician agreement to patient requests for prayer behaviors. Unmarried physicians and family practitioners tended to agree to patient requests for participating in prayer in every setting, while only marital status held the same association for silent prayer in the hospital setting. No independent association was noted for the number of years in practice, resident status, age, or sex with any of these outcomes. When SWB score was used as a continuous rather than a trichotomous variable, similar results were obtained.
Primary care physician attitudes toward involvement in patients' religious and spiritual life are heavily dependent on the clinical setting and the specific behavior involved. As previously reported,20- 22 our survey confirms that, when asked if they should be aware of patients' religious and spiritual beliefs, most physicians agree. When asked about specific behavior in varied clinical settings, however, most thought they should inquire about such beliefs only in the setting of a dying patient. This reluctance to initiate involvement in patients' religious and spiritual life has been suggested in prior studies. Survey data confirm that, despite agreeing that they should be aware of patients' spiritual history, physicians are not acting on these stated beliefs.13,14,21,22,25,26 Reasons for avoiding the topic have been proposed and include lack of time, lack of training in how to obtain a spiritual history, difficulty in identifying receptive patients, concerns about projecting beliefs onto patients, and uncertainty on how to address spiritual issues raised by patients.22 Given the response to patient requests for spiritual behavior in our study, the most significant barriers to physician adaptation of such behavior appear to be the difficulties of identifying receptive patients and concerns about projecting beliefs onto patients.
No matter the intervention (spiritual history, silent prayer, or active prayer), physicians are more likely to agree to participate as the severity of the patient's illness increases. Still, the only behaviors receiving majority agreement were asking dying patients about their religion and spirituality (74.2%) and saying a silent prayer for dying patients (54.1%). In a small study25 of devout physicians, physician awareness of patients' beliefs and physician prayer for patients were also noted to increase with increased severity of illness. Koenig et al20 found that 65.8% of family practitioners considered it appropriate to address religious and spiritual issues with bereavement or death even if the patient gives no indication that religion is important to him or her. In a study14 of Vermont family physicians, religious inquiries were noted most for terminal illness (69%) and near-death (68%) counseling. Furthermore, Ellis et al22 noted that fear of death and dying was the spiritual topic most often discussed by physicians, especially in the hospital or nursing home. Perhaps as the severity of a patient's illness increases, physicians believe the relative importance of his or her religious and spiritual beliefs increases and therefore are more comfortable addressing them directly.
Although each spiritual behavior was supported more readily as the severity of the clinical setting worsened, the more intense the spiritual behavior, the less physicians agreed to its appropriateness. For a dying patient, only slightly more than one third of those willing to take a spiritual history would initiate prayer with the patient. This reluctance to engage in more demonstrative spiritual behaviors may be due to the multiple factors noted by Ellis et al.22 Perception of patient desires might also play a role. In a survey of primary care physicians, Koenig et al20 found that 62.7% thought patients would not want physicians to pray with them even in severe illness or emotional distress. Furthermore, physicians may believe that they are stepping beyond their professional role with more active spiritual behaviors or, more strongly, that it is unethical to engage patients without a specific patient request.4,27
In our study, despite a general reluctance to be involved in patients' religious and spiritual life, physicians appeared to be willing to fulfill patient requests regarding their religious and spiritual needs. With specific patient request, most physicians agreed to silent prayer for or prayer with patients in all clinical settings. The gap between the percentages of physicians willing to perform different intensities of prayer behavior for each clinical setting nearly disappeared when patient request was added to the mix. This finding was previously suggested in a study20 of Illinois family practitioners in which 87.7% agreed that physician involvement is appropriate when the patient makes a direct request for help. Again, a likely interpretation of this finding is that, while many physicians recognize the importance of such beliefs to patients, they are not involved because they have trouble identifying receptive ones or do not want to think they are pushing a faith on them. It may also indicate a level of discomfort in not addressing a need identified directly by a patient during the medical encounter. We did not ask physicians if they would refer a patient with a prayer request to a chaplain or religious leader if they were unwilling to comply.
In regression analysis, we found increased agreement of family practitioners compared with internists toward inquiry into religious and spiritual matters—essentially taking a spiritual history. When more active spiritual behaviors were entertained, this difference disappeared. Family practice training in spirituality and psychosocial issues might explain this difference. However, only religious spiritual fervor (not taught in residency) seems to drive physicians to initiate more intense spiritual behavior on their own. As noted in a prior study25 of devout physicians, even in this group the incidence of physician prayer with patients is low. Regression analysis further noted a pattern of association with unmarried physicians being more willing to be involved in certain prayer behaviors, although the reason for this finding is not entirely clear.
Limitations of this study include the possibility that nonrespondents differed from respondents. Despite numerous attempts to elicit participation, slightly more than one third of physicians surveyed did not respond. This is similar to the response rates of other physician surveys of this size.21,28 Also, as we focused on a limited geographic area, with most physicians surveyed practicing in North Carolina, the findings may not relate to all US physicians. However, in our geographic distribution, the mean ± SD SWB score for all respondents was 93.4 ± 16.6, which is lower than the 100.7 ± 16.5 reported for Missouri family physicians.22 This comparison tempers the idea that the physicians in this study are unusually religious compared with other parts of the nation.
In conclusion, most primary care physicians agree to spontaneous spiritual involvement with their patients if this involvement is limited to a religious and spiritual history and to the setting of death and dying. A more vigorous spiritual role is unlikely to be physician initiated but can be instigated by patient requests. Although studies of patient attitudes have consistently shown a desire for more attention to the religious and spiritual aspects of illness and health, it appears unlikely that the true medical benefit of spiritual interventions, especially prayer, will ever be clearly defined.27,29 Therefore, before making a blanket admonition for patients to request such involvement from their physicians, it would be prudent to recommend further research to identify more realistic end points to measure spiritual care (eg, satisfaction and quality-of-life measures) and the extent to which physicians, as opposed to clergy, counselors, family, or friends, are the appropriate vehicle to achieve these end points.
Corresponding author and reprints: Michael H. Monroe, MD, Department of Internal Medicine, Carolinas Medical Center, 1000 Blythe Blvd, Charlotte, NC 28203 (e-mail: email@example.com).
Accepted for publication January 24, 2003.
We thank our collaborator Andria Klioze, MD, and the faculty and staff of The University of North Carolina at Chapel Hill General Internal Medicine Faculty Development Fellowship Program.
This study was supported by The University of North Carolina at Chapel Hill General Internal Medicine Faculty Development Fellowship Program, which is funded by grant 5-D08-HP54004 from the Health Resources and Services Administration, Rockville, Md. Additional support was provided by grants from the Charlotte-Mecklenburg Health Services Foundation, Inc, the University of Florida College of Medicine Education Center, and the Moses Cone Medical Education Research Committee.