- © 2003 by American Society of Clinical Oncology
Importance of Faith on Medical Decisions Regarding Cancer Care
- From the Department of Medicine and Center for Health Care Research, Medical University of South Carolina, Charleston, SC.
- Address reprint requests to Gerard A. Silvestri, MD, MS, Associate Professor of Medicine, Medical University of South Carolina, 96 Jonathan Lucas St, Ste 812 CSB, P.O. Box 250623, Charleston, SC 29425; email: Silvestri{at}musc.edu.
Abstract
Purpose: Decisions regarding cancer treatment choices can be difficult. Several factors may influence the decision to undergo treatment. One poorly understood factor is the influence of a patient’s faith on how they make medical decisions. We compared the importance of faith on treatment decisions among doctors, patients, and patient caregivers.
Methods: One hundred patients with advanced lung cancer, their caregivers, and 257 medical oncologists were interviewed. Participants were asked to rank the importance of the following factors that might influence treatment decisions: cancer doctor’s recommendation, faith in God, ability of treatment to cure disease, side effects, family doctor’s recommendation, spouse’s recommendation, and children’s recommendation.
Results: All three groups ranked the oncologist’s recommendation as most important. Patients and caregivers ranked faith in God second, whereas physicians placed it last (P < .0001). Patients who placed a high priority on faith in God had less formal education (P < .0001).
Conclusion: Patients and caregivers agree on the factors that are important in deciding treatment for advanced lung cancer but differ substantially from doctors. All agree that the oncologist’s recommendation is most important. This is the first study to demonstrate that, for some, faith is an important factor in medical decision making, more so than even the efficacy of treatment. If faith plays an important role in how some patients decide treatment, and physicians do not account for it, the decision-making process may be unsatisfactory to all involved. Future studies should clarify how faith influences individual decisions regarding treatment.
MEDICAL DECISIONS can be complicated even in the best of circumstances. However, when a patient is faced with a life-threatening illness, deciding which treatment option is best can become extremely difficult for both the patient and doctor. Nearly 60,000 patients per year in the United States will present with newly diagnosed metastatic lung cancer, a disease associated with little hope of long-term survival. The treatment options are chemotherapy or supportive care. Chemotherapy can produce improvements in quality of life and adds a small survival benefit. We have previously studied the decision-making process for patients with advanced non–small-cell lung cancer who had been treated with chemotherapy.1 In that study, we presented scenarios to patients that described the side effects and likely survival benefit of chemotherapy. Our results indicated that the majority of patients would have selected supportive care instead of chemotherapy. Because that choice appeared incongruous with the treatment patients received, we wondered which factors might influence patients’ treatment decisions.
There has been a recent interest in the role of religion in medicine.2–,10 Most of this literature compares outcomes of patients who are regular churchgoers with those who are not. Some studies document a better health outcome in patients with a strong religious faith.6 There has been vigorous debate about whether the results of these studies are valid and whether physicians should even address the issue of religion with patients.9–,12 Few studies explore the role that faith plays in medical decision making.13
As part of a larger study on decision making and lung cancer, we chose to assess the relative influence of faith on the decision-making process for patients with newly diagnosed advanced or metastatic lung cancer. We compared their responses with those of their primary caregivers and with those from a group of medical oncologists.
METHODS
The study consisted of three surveys, each to one of the following different populations: patients with newly diagnosed advanced (stage IIIB/IV) lung cancer, their caregivers, and medical oncologists. One hundred new patients seen at two cancer clinics (the Medical University of South Carolina’s Hollings Cancer Clinic and the Department of Veterans Affairs Medical Center, Charleston, SC) were invited to participate in this study, as were their respective caregivers. Patients and caregivers were interviewed separately after they had been informed of the patient’s diagnosis. A trained, noninvestigator interviewer administered the survey to the patients and caregivers in the clinic immediately after a physician visit.
A convenience sample of medical oncologists attending the Thirty-Fifth Annual American Society of Clinical Oncology Meeting (May 1999) was selected to participate in the survey. Two project staff assistants attended the American Society of Clinical Oncology annual meeting as surveyors and were stationed in the American Cancer Society booth during 3 days of the meeting. They were instructed to offer the printed survey to any meeting attendant passing by the exhibit booth. The surveyor used a written script identifying the survey sponsor, investigator institution, and survey purpose. Participants were asked to complete the survey at time of receipt and return it to the surveyor. Responses from all completed surveys were entered into a computerized database (data were double entered for accuracy) and exported to SPSS (SPSS, Inc, Chicago, IL) and SAS (SAS Institute, Cary, NC) statistical software for analysis. The institutional review board of the Medical University of South Carolina granted approval for the study.
Questionnaire Design
The patient questionnaire consisted of demographic questions such as age, race, sex, marital status, insurance coverage, and education level. Clinical information regarding disease stage, cell type, and the patients’ performance status was obtained for all patients from clinic medical records. All patients were surveyed after the diagnosis was established and before the decision regarding treatment and subsequent care was delivered. All patients were facing the decision of chemotherapy versus supportive care for treatment of their metastatic lung cancer.
Participants were asked to rank how the following factors would influence their choice between two different treatment options: (1) ability of the treatment to cure the disease; (2) family doctor’s recommendation; (3) medical oncologist’s recommendation; (4) spouse’s recommendation; (5) children’s recommendation; (6) faith in God; and (7) side effects of treatment.
Caregivers and physicians were asked to rank these same factors in terms of which factor they thought should influence the patient’s decision for treatment. Patients and caregivers ranked factors using the Thurstone methodology, which is described in detail elsewhere.14,15 Briefly, the Thurstone method of ranking has the respondent view two of the choices shown on separate cards and asks them to pick the choice that would most influence their decision regarding treatment.
The process is repeated with every different combination of choices. From that, a rank order is established. Because of time constraints the medical oncologists were asked to use a simple rank order to rank their choices from 1 to 7. A group of 50 physicians were asked to rank the factors using either the Thurstone method or simple rank order. A week later they were asked to rank the factors using the alternative ranking method. Analysis of this group showed no difference in rank order by the different methods.
Statistical Analysis
Rankings were compared among the three groups using restricted simultaneous ordered logistic regression models. Using this method accounted for the inherent dependence among the seven items’ rankings. For example, the models accounted for the fact that if a respondent ranked item 1 as the most important, none of the other six items could be ranked as most important. From the models, statistical tests were performed that determined whether there were any significant differences with respect to overall rankings between patients and caregivers, between patients and physicians, and between caregivers and physicians. In addition, the models allowed us to assess whether there were significant differences in the rankings of individual items. Comparisons between those patients who ranked faith as a priority (rank, 1 or 2) and those who ranked faith as a low priority (rank, 3 to 7) were made using t tests and χ2 tests as appropriate. In addition, a logistic regression model was used to determine which patient characteristics were significantly and independently associated with ranking faith high or low.
RESULTS
One hundred patients with advanced lung cancer and their caregivers were interviewed separately for this study. The mean age of the patients was 61 years. Seventy-two percent were male, and 30% were African-American. Nearly 60% had at least a high school degree. Fifty-nine percent were married. The majority of patients (70%) had metastatic (stage IV) non–small-cell lung cancer, with the remainder divided between stage IIIB (15%) and IIIA (9%) non–small-cell lung cancer and limited (1%) and extensive (5%) small-cell lung cancer. Patients presented with a good performance status (mean Karnofsky performance status, 77.9). Table 1⇓ lists the physicians’ characteristics, including the distribution of the physicians’ rankings of the importance of faith in God in making treatment decisions.
Physician Characteristics (N = 257)
The rank order of factors that patients, caregivers, and oncologists used to decide treatment is listed in Table 2⇓, with a summary of the logistic regression models presented in Table 3⇓. Overall, patients’ and caregivers’ rankings were not significantly different from one another, whereas both patients’ and caregivers’ rankings differed substantially (P < .0001) from those of the oncologists. In fact, patients and caregivers were in complete agreement with the average rankings of the factors that would influence decision making for their cancer care. All three groups also agree that the oncologists’ recommendation regarding treatment was the single most important factor. The beta coefficients presented in Table 3⇓ reflect the degree to which each item influenced the summary statistic, with positive beta coefficients indicating that the first group listed ranked the factor of interest higher, on average, than the second group. Thus, when patients and physicians are compared, the largest discrepancy in the factor rankings was the difference in the rankings for faith in God, as evidenced by the beta coefficient for faith in God (beta = 2.42) being larger in absolute value than any of the others. Likewise, when caregivers and physicians are compared, the largest beta coefficient was that for the faith in God factor (beta = 2.24), reflecting the disparity between the two groups’ ranking of this factor. Because patients and caregivers ranked faith in God so differently from oncologists, the rankings for other factors were influenced as well, making it appear, for example, that oncologists also felt that the spouses’ input should be more influential than did either patients or caregivers. Among oncologists, there was no difference in rank order when physician characteristics were compared (eg, country of residence, age of physician, practice setting, number of patients treated per year, and so on). The differences between patients who ranked faith high (rank, 1 or 2) versus those who ranked faith low (rank, 3 to 7) are listed in Table 4⇓. In bivariate (unadjusted) analyses, race and education seemed to be significantly associated with faith. However, when the patients’ age, sex, race, and education were included in a logistic model to measure each variable’s independent influence on faith, the only variable that remained significant (P < .05) was education, with more educated patients being less likely to rank faith as high.
Patient, Caregiver, and Physician Rankings of Treatment Decision Factors
Beta Coefficients Obtained From the Restricted Simultaneous Ordered Logistic Regression Models
Differences Between Patients Who Ranked Faith in God High (1 or 2) and Those Who Ranked It Lower (3 to 7)
DISCUSSION
This study found that when deciding between different treatment options, faith in God seems to be an important factor for patients with newly diagnosed advanced lung cancer and their caregivers. It ranked second only to the recommendation of the medical oncologist, which everyone agreed was the most important influence in the decision process. Patients and caregivers tended to weigh faith more importantly than even the ability of treatment to cure the disease. Conversely, physicians felt a patient’s faith in God should be the least important factor that patients should consider when deciding therapy.
Recently, there has been a resurgence of interest in the relationship between religion and medicine. It has been reported that 84% of adult Americans believe that God performs miracles, and 48% say that they have witnessed one.16 There has also been research on this subject directed at the effect of faith or a high level of spirituality on outcomes of medical illnesses. Luskin6 performed an extensive review of the effect of spiritual and religious factors on morbidity and mortality, with a focus on cardiovascular and pulmonary disease. The majority of studies reviewed revealed evidence of improved health outcomes for those who are regular churchgoers as opposed to those who are not. Some of these studies concluded that there may be some unexplained effect of faith, religion, or God that improves medical outcome.4 These studies have been intensely critiqued for design flaws. Critics of this research point out that regular churchgoers may live healthier lifestyles (ie, nonsmokers or alcohol users) as opposed to nonchurchgoers. The failure to control for these possible confounding variables could explain the measured differences in health outcomes.9,11
Relatively little has been written about the role of religion in medical decision making. Ehman et al13 queried 177 patients with end-stage chronic obstructive pulmonary disease and asked them whether faith played a role in deciding end-of-life care issues. They found that 45% of patients felt that faith would play a role in decision making and that, if tactfully approached, 80% of patients would not mind being asked about religion.
How do these findings apply to the role of faith in decision making? The controversy about the effect of faith on health outcome has led some to oppose any discussion of religion between health providers and patients. Sloan et al10 go further to state that “the mere fact that patients want to talk about religion doesn’t mean it should occur.” If such discussions do not take place for certain people, then faith may not be accounted for when patients decide treatment. This could lead to some patients being dissatisfied with their decisions regarding treatment and could ultimately hurt the physician-patient relationship.
Why do patients rely so heavily on faith when deciding treatments? It is surprising that patients say faith will be more influential in their decision for treatment than the ability of treatment to cure the disease. Those who rated faith high (rank, 1 or 2) were less educated even after adjusting for other potential confounders. One possible explanation for the fact that less educated patients tended to rank faith high may be that they do not understand the technical aspects of treatment options, such as differing chemotherapy regimens and the side effects of those regimens. They may not understand survival curves or median survival benefit. They may resort to religion, a concept that does not necessarily require a strong educational background but does require faith. It is also possible that many patients discover religion when faced with a serious illness. Roberts et al8 reported that, in 108 women with gynecologic cancers, 76% stated that religion had a strong place in their lives, and nearly half of those women (49%) felt that they had become even more religious since the diagnosis of cancer was made. Perhaps in the end, faith in God provides patients with hope, a concept that is commonly referred to in cancer care but that is not easily quantified.
Why do doctors rank faith as the factor that should least influence patients when deciding treatments? This study provides no clear answers to that question. It is unlikely that we surveyed 257 oncologists who have no personal religious beliefs. The survey included physicians from 42 countries representing diverse religious influences, who, as a group, tended to rank faith last. However, a subset of that group, nearly 15%, did rank faith as either the most or next most important influence patients should consider when deciding treatment. Others have studied the religious beliefs of physicians and found both geographic and specialty variation in the religious convictions of physicians.17,18 A national survey of family physicians found that their religious practices mirrored those of the general public.17 Conversely, psychiatrists as a whole were much less likely to believe in God or to attend church as regularly as the general public.19
One can postulate that physicians rank faith low because their training is based so heavily on scientific reason that they are unable to recommend a concept that defies a reasonable explanation. It may also simply reflect discomfort in exploring such a personal issue.
These findings have implications for medical decision making. Formal decision analysis imputes the probabilities of certain outcomes with the utilities for those outcomes into a model where a choice between two treatment options is computed. The decision presented here is essentially supportive care versus chemotherapy for the treatment of metastatic lung cancer. Other decision models have examined this subject and used utilities such as survival, quality of life, cost, and side effects of therapy. This study indicates that this type of decision model may be overly simplistic and that patients who value faith more than the aforementioned tangible probabilities would not have that reflected in the modeling processes.20
The major limitation of this study is that all patients and caregivers interviewed were from the same geographic location, namely the southeastern United States. This region has traditionally been considered part of the Bible Belt, an area of the United States in which many of the people are known to have particularly strong religious beliefs, and the responses our patients gave us may not represent those found elsewhere. However, other studies from geographic locations throughout the world have also found a high level of religious belief among patients.21,22
In summary, to our knowledge, this is the first study that shows a marked difference in how patients, their caregivers, and physicians view the influence of faith in medical decision making. Patients and their caregivers seem to rely heavily on their faith to help them decide on treatment for their cancer. It seems that physicians may not recognize the importance patients place on religion when faced with a life-threatening illness. These findings do not shed light on how or even whether this subject should be handled when it arises in discussions between physicians and patients. At the very least, it seems that physicians should not discount the strongly held beliefs that patients may have. Acknowledgment and respect by physicians of a patient’s personal beliefs will likely lead to higher satisfaction with the decision-making process for all involved.
Footnotes
-
Supported by grant ACS IRG 97-151-01 from the American Cancer Society.
- Received August 6, 2002.
- Accepted December 12, 2002.
