VIEW CONTENTS OF PREVIOUS ISSUES

Autonomy, Gender, and Preference for Paternalistic or Informative Physicians: A Study of the Doctor-Patient Relation

Kalman J Kaplan, PhD, Mark Schneiderhan, PhD, Martin Harrow, PhD, and Russell Omens, PsyD

In the past few decades, medical ethicists^1,2,3 have departed from the Hippocratic model in encouraging patients to take a more active role in their healthcare.^4,5,6 Emanuel and Emanuel⁴ contrast informative versus paternalistic styles as endpoints in a continuum of doctor-patient relationships. In the paternalistic model the physician acts as the patient¹s guardian and articulates what is best for the patient.

In this model, the physician-patient interaction ensures that patients receive interventions that best promote their health and well-being. To this end, physicians use their skill to determine the patient¹s medical condition and his or her stage in the disease process and to identify the medical tests and treatments most likely to restore the patient¹s health or ameliorate pain. Then the physician presents the patient with selected information that will encourage the patient to consent to the intervention the physician considers best. At the extreme, the physician authoritatively informs the patient when the intervention will be initiated. (p. 2221)

The paternalistic model assumes that there are shared objective criteria for determining what is best for the patient. Hence, the clinician can discern with limited patient participation what is best for the patient.

In the informative model, the physician treats the patient as a consumer.

The objective of the physician-patient interaction is for the physician to provide the patient with all relevant information, for the patient to select the medical interventions he or she wants and for the physician to execute the selected interventions. To this end, the physician informs the patient of his or her disease state, the nature of possible diagnostic and therapeutic interventions, the nature and probability of risks and benefits associated with the interventions and any uncertainties of knowledge. It is the physician¹s obligation to provide all of the facts, and the patient's values then determine what treatments are to be given. (p. 2221)

In the informative model, the physician provides the patient the means to exercise control and to make an informed decision as to the course of treatment.

The present study explores the relationship of physician communication style to patient characteristics. Do some patients prefer one type of physician style and other patients another? Do some patients need the freedom implicit with an informative physician style while other patients need the protection implicit in the paternalistic physician style?

Some studies^7,8,9 report that younger and higher educated individuals are more likely to take an active role in medical decisions while men, married individuals, and patients with a more severe prognosis tend to prefer a more passive role and allow the physician to make medical decisions.¹⁰ However, another study¹¹ reported that patient gender accounts for only a small proportion of the variance of decision-making style.

Several recent studies have specifically examined the decision-making style and preferences of patients diagnosed with cancer. One study¹² reports that an index group of 150 patients newly diagnosed with breast cancer were more likely to want to play a passive role in decision-making as compared to 200 women with benign breast disease. In both the index and control groups, older women and women of lower social class preferred a more passive role. A second larger study¹³ examined 1,012 women with breast cancer. Women younger than 50, married, with English as a first language, having greater than a high school education, at earlier stages of the disease, and having undergone a lumpectomy were more likely to prefer active or collaborative roles in decision-making. The third study in this series studied examined men diagnosed with prostate cancer.¹⁴ This study reports that among 60 men newly diagnosed with prostate cancer, rehearsal in self-efficacy tended to increase active participation in treatment decisions.

Unfortunately, none of these three studies take into account physician style4 per se nor did they measure relevant patient personality characteristics such as locus of control or autonomy. Finally, each of these studies was nested within gender making it impossible to compare differences in gender in this regard. This is especially unfortunate because of the literature on gender differences in help-seeking behavior. A considerable amount of work, for example, indicates that women are more open than men to receive and utilize social supports^15-19 though one study²⁰ reports that gender differences are a function of sex role characteristics rather than sex per se.

The present study attempts to link this body of literature with the Emanuel and Emanuel model.⁴ We hypothesize that non-autonomous individuals prefer paternalistic physician styles and autonomous patients, informative physician styles. However, this pattern may be affected by patient gender. The present study examines patient preference in clinician styles as a function of general patient autonomy level, knowledge of the specific disease, and patient gender.

Method
Sample
One hundred thirty-one ambulatory care patients at Michael Reese Hospital and the University of Illinois at Chicago Medical Center were surveyed. The study was approved by the Institutional Review Boards of Michael Reese Hospital and the University of Illinois Medical Center. Of the 131 patients, 91 were women and 40 were men. The patients were of the following ethnic backgrounds: 18 Caucasian, 54 African American, 21 Asian American, 3 Hispanic, 32 other, and three unknown. The average age of the men was 37.6 years and the women was 45.6 years. Thirty-eight percent completed high school and 61% completed college. Thirty percent were professional, 20% were clerical, and 46% were unemployed.

Survey overview:
The survey was comprised of three major parts. The survey took approximately 15 ­ 20 minutes to complete. The written survey was administered in the clinic waiting room by a research associate. The research associate was available to answer any questions pertaining to the study and to read the survey to the subject, if necessary, in a private location to ensure confidentiality. The first part was presented to all subjects and asked general demographic information about the subject. The second and third parts contained scenarios and questions specific to gender. In the second part, male subjects were asked to imagine that they were recently diagnosed with prostate cancer and were about to see a specialist to determine the treatment plan. Likewise, female subjects were asked to imagine that they had been recently diagnosed with breast cancer. Both male and female subjects were then presented with two potential physician response styles adjusted for the gender-specific disease: Dr. P (paternalistic style) and Dr. I (informative style). The third part measured the subjects¹ autonomy level on the Individuation-Attachment Questionnaire (IAQ). Finally, subjects were asked questions about their degree of knowledge of the specific imagined disease state (prostate cancer for men and breast cancer for women).

Independent variables:
The major independent variables in this study consisted of the following: a) the degree of subject autonomy (high versus low), b) subject knowledge of the disease, c) subject gender, and d) the physician response style (paternalistic or informative).

Subject autonomy level. Subject autonomy level was conceptualized as an overall personality style and was measured on the (IAQ)^21,22,23 which has been used in the past in connection with NIMH psychological autopsy study on adolescent suicide²⁴ and consists of twenty five-point Likert scales asking subjects to agree or disagree with a series of statements dealing with attachment and individuation themes. It is designed to yield four separate attachment and individuation scores: Need for Individuation (NI) (the need to make one's own decisions: e.g., "I believe everyone must find his or her own way in life."), Fear of Individuation (FI) (the fear of making one's own decisions: e.g., "It is important for me to do what other people think I should do."), Need for Attachment (NA) (the need to form close relationships: e.g., "I need to share my feelings with others."), and Fear of Attachment (FA) (the fear of forming close relationships: e.g., "A close relationship makes it hard to be yourself."). It has been employed in a number of studies involving over 2,000 respondents at various ages across the life-span, both clinical and non-clinical, both male and female. Reliabilities on the four sub-scales ranged from .75 (FI) to .84 (NA) and inter-correlations between the four sub-scales are quite low (.22 to -.19). Validity coefficients of the four scales with independent clinical assessments of these same characteristics ranged from .58 (FI) to .71 (NI).

Subject knowledge of disease. Subject knowledge, of the specific disease state they were asked to imagine being diagnosed with, was measured on four three-point scales, varying from a little to a lot:

a) Overall Familiarity (OF): To what extent have you heard about breast (prostate) cancer?
b) Understanding of Effects (UE): Do you understand the effects of breast (prostate) cancer on a person's life?
c) Understanding of Treatment Options (UTxO): Do you understand the treatment options for breast (prostate) cancer?
d) Understanding of Risks/Benefits (UR/B): Do you understand the benefits and risks of the different treatments?

Subject gender. Subject gender is self-explanatory. Men were presented with the following scenario:

Try to imagine being confronted with the following circumstance: You started noticing symptoms for prostate cancer: difficulty urinating and a nagging pain in the hips. These symptoms lead you to see an oncologist, who runs a series of tests. The two paragraphs below describe different hypothetical doctor responses to your situation. Please read them both and indicate which way you would prefer your doctor to act.

Women were presented with this alternate scenario:

Try to imagine being confronted with the following circumstance: You started noticing symptoms for breast cancer: skin irritation and a small lump around the breast area. These symptoms lead you to see an oncologist, who runs a series of tests. The two paragraphs below describe different hypothetical doctor responses to your situation. Please read them both and indicate which way you would prefer your doctor to act.

Physician response style (Paternalistic-Doctor P and Informative-Doctor I) is manipulated separately for male and female subjects. The styles for the male subjects are presented as follows:

Doctor P: The last time we met, I told you that we found a tumor and that the cancer cells are found only in the prostate gland. I know that you are probably afraid and anxious about what having the cancer means. The news about the cancer must be shocking to you and speaking to you in our last visit, I know that you have had the support from your family and friends to cope with the disease. Given the stage of your cancer, the results of the various tests, and your medical history, you are a very good candidate for surgery. The sooner we schedule an appointment for the surgery the better. I am here to take care of you and to see you through this thing.

Doctor I: The last time we met, I talked to you about prostate cancer and briefly introduced treatments. We know from the tests that the cancer cells are localized in the prostate gland, and they have not yet spread to the other tissues surrounding the gland. For this stage of the disease, we can perform surgery, what we label in medical terms -a radical prostatectomy, to remove the prostate and some of the tissue around it. Unfortunately, the side effects of this operation are impotence and leakage of urine from the bladder. The other options outside of surgery are external beam irradiation or interstitial radioisotopes, which offer similar therapeutic effects to a prostatecomy. This therapy can lead to impotence and other side effects. If you are not too clear about the treatments, we have some pamphlets that will summarize what I have just told you.

The two styles for the female patients are as follows:

Doctor P: The tests that we performed during the last visit indicate that you have second stage breast cancer. The cancer is localized, and the tests show that the cancer has not yet spread to other parts of your body. A lumpectomy combined with radiation therapy offers the greatest chance for survival and the most favorable cosmetic result. Right now we need to schedule an appointment for the surgery as soon as possible. I have seen many patients with your type of breast cancer who have undergone these same procedures, and they have had good outcomes. I am here to take care of you and to see you through this thing.

Doctor I: The last time we met, I talked to you about breast cancer and briefly introduced treatments. Now that you have had some time to cope with this disease, we need to talk more about the treatment options. There are two main issues involved: local and systemic control. With local control, the options are to perform a lumpectomy or mastectomy with or without radiation. The lumpectomy removes only the cancer and some of the surrounding tissue, while the mastectomy can remove a segment or the whole breast. The studies have shown that mastectomy and lumpectomy combined with radiation result in identical overall survival, about an 80% survival of ten years. Lumpectomy without radiation results in a 30-40% chance of tumor recurrence in the breast. If you are not too clear about the treatments, we have some pamphlets that will summarize what I have just told you.

In addition, other demographic information such as age, job, ethnic background, marital and family status, and education was collected as well as more particular information regarding knowledge, of the disease they were asked to imagine they had, as well as anticipated family support.

Dependent variables: The subjects were asked: a) to decide which physician response style (Dr. P or Dr. I) they would want their own physician to act like and to describe why; and b) to rate Dr. P and Dr. I on a series of 9 seven-point semantic differential scales (i.e., nonhelpful-helpful, disrespectful-respectful, uninformative-informative, nonsupportive-supportive, aloof-compassionate, nondirective-directive, nondemocratic-democratic, nonprotective-protective, and nonreceptive-receptive).

Results
Autonomy and Knowledge of Disease.
The intercorrelations between measures of subject autonomy and subject knowledge of either prostate cancer (for men) and breast cancer (for women) are presented in Table 1. We present this data in three sections: a) the inter-correlations between measures of subject autonomy, b) the inter-correlations between measures of subject knowledge of disease, and c) the correlations between autonomy measures and measures of knowledge of disease.

*p<.05, Ýp<.01 NI=Need for Individuation, FI=Fear of Individuation, NA=Need for Attachment, FA=Fear of Attachment, OF=Overall Familiarity with Disease, UE=Understanding of Effects of Disease, UTxO=Understanding of Treatment Options, UR/B=Understanding of Risks/Benefits of Treatment

Intercorrelations of Autonomy Measures. The four scales used to assess subjects¹ overall autonomy shows markedly different patterns for men and for women. NI, our basic index of autonomy, is negatively related to FA for male subjects (r=-.69, p<.05) but positively related to this same variable for female subjects (r=.43, p<.05). NI is unrelated to FI for men (r=-.09, n.s.) but is positively related to FI for women (r=.43, p<.05). Finally, NI is very highly correlated with NA for men (r=.71, p<.01) but not significantly so for women (r=.32, n.s.). This pattern suggests that autonomy may function differently for men and women in our sample. For male patients, autonomy is unrelated to either FA or FI and indeed is positively related to NA. For female patients, however, it does seem to be intertwined with both FA and FI and unrelated to NA.

Intercorrelations of Knowledge of Disease Measures. The four scales used to assess subjects' knowledge of disease were highly intercorrelated for both men and women. The sizes of these correlation ranged from .37 to .81 (p's<.01).

Correlations Between Measures of Autonomy and Knowledge of Disease. Very few of the correlations between the four autonomy scores and the four knowledge scores are significant. However, the data reveals tendencies for positive relations between the autonomy (NI) and knowledge scores for men (r's=.17, .28, .26, .23) and tendencies for negative relations between autonomy and knowledge for women (r's=-.17, -.31Ý, -.23*, -.13, Ý=p<.01, *=p<.05). Much of the same results can be seen in chi-square analyses of the relationship between dichotomized indices of autonomy and knowledge of disease. No relationship emerged between these two variables for males (chi-square=.65, n.s.). For females, in contrast however, a strong negative relationship emerged. Seventy-three percent of low autonomy women reported high knowledge of the disease as compared to 45% of high autonomy women (chi-square=4.5, p<.03). This is somewhat understandable given the "fear of attachment" aspect of the female autonomy index. Acquiring knowledge from others about a disease may be predicated on an openness toward interaction with them (i.e., overcoming one's fear of attachment). For the sake of simplicity, all further analyses in this paper will employ the NI index of autonomy and OF, the most omnibus measure of knowledge of disease: "To what extent have you heard about breast (prostate) cancer?"

Autonomy, Knowledge of Disease and Physician Preference
Subject Level of Autonomy and Preferred Physician Style. Males and females showed no difference in preference for physician style, with 51% of the male patients and 55% of the female patients preferring the informative style (chi-square=.11, n.s.). Further, males and females did not significantly differ from each other in autonomy levels (t=1.80, n.s.). However, the interaction between gender and autonomy level was quite unexpected. Overall, 66% of highly autonomous (upper third) patients preferred an informative clinician style as compared to 45% of non-autonomous (lower third) patients (chi-square=3.57, p=<.06). Although males and females did not differ in autonomy levels, autonomy was linked to doctor style preference only in the female patients. Among female patients, 68% of autonomous patients preferred an informative clinician style as compared to only 41% of non-autonomous female patients (chi-square=4.25, p<.04). No difference emerged for male patients, with 60% of autonomous patients and 53% of non-autonomous patients (chi-square=.11, n.s.) preferring an informative clinician style.

Subject Knowledge of Disease and Preferred Physician Style:
The results for overall knowledge of the disease are both dissimilar and similar to those involving autonomy. Female subjects report having a greater familiarity with breast cancer (1.59) than do men with regard to prostate cancer (1.29, t=3.20, p<.01). This pattern is different than that regarding autonomy where no significant differences were reported between men and women. Nevertheless, the relationships between knowledge of disease and preferred physician style are quite similar to those involving autonomy.

Overall, 66% of people who reported high familiarity with their imagined disease preferred an informative physician style as compared to 47% of people who reported moderate or low familiarity (chi-square=4.26, p<.05).

The interaction between gender and knowledge of disease was similar to the pattern described above regarding autonomy. Among female patients, 66% of patients with high familiarity with their imagined disease preferred an informative style as compared to only 44% of patients with low familiarity (chi-square=3.70, p=.05). The difference among the male patients in this regard was not significant (chi-square=.74, n.s.). In other words, preference for an informative physician style increased with familiarity with their imagined disease for female but not for male subjects. (refer to figure 2)

Level of Autonomy, Knowledge of Disease, and Preferred Physician Style for Men and Women. This section presents chi-square data on the three-way relationship between autonomy, knowledge, and physician preference for men and women. Among low autonomy women, knowledge of disease is significantly positively related to preference for an informative physician (chi-square=3.96, p<.05). Fifty-nine percent of low autonomous women with higher knowledge of breast cancer preferred informative physicians as compared to only 14% of low autonomy women with some or little knowledge. The same trend (greater knowledge of breast cancer associated with informative physician styles) emerged for high autonomy women, but the relationship was not significant (chi-square=2.31, n.s.). The same pattern can be looked at in the opposite way as well. For women with low knowledge of breast cancer, high autonomy is linked to preference for informative physician style (chi-square=3.49, p=.06). The same trend (greater autonomy associated with preference for informative physician styles) can be seen among women with high knowledge of breast cancer as well, but once again the relationship was not significant (chi-square=.24, n.s.). Among men, no such relationships emerged.

What this data seems to indicate is that, for women, autonomy is most important in predicting physician preference when knowledge of the disease is low or moderate. Likewise, knowledge of the disease is most important for predicting physician preference when autonomy is low.

Beliefs about Physicians and Physician Preference
This section examines the relationship between beliefs about physicians and physician preference. In these analyses, a positive relationship indicates that the preferred physician (either paternalistic or informative) was more positively evaluated on the nine adjectival scales than the non-preferred physician. The multiple correlation between discrepancies in all nine predictor scales (i.e., nonhelpful-helpful, disrespectful-respectful, uninformative-informative, nonsupportive-supportive, aloofness-compassionate, nondirective-directive, nondemocratic-democratic, nonprotective-protective, and nonreceptive-receptive) was .88 for male subjects and .82 for female subjects (p<.01 in both cases).

Male patients' physician preferences were most influenced by discrepancies in perception of clinician helpfulness (r=.84, p<.01) and respectfulness (r=.81, p<.01) and least influenced by discrepancies in perceived clinician directiveness (r=.15, n.s.). Female preferences of clinician style were influenced by slightly different factors. They were most influenced by discrepancies in clinician helpfulness (r=.76, p<.01) and supportiveness (r=.73, p<.01) and least influenced by discrepancies in perception of clinician compassion (r=.16, n.s.). Both patient autonomy and patient knowledge of his or her imagined disease affected the relative importance of discrepancy in perceived clinician informativeness for physician preference. Informativeness was the second most important attribute after helpfulness for low autonomy men (r=.73, p<.01) but the least important attribute for low autonomy women (r=.38, n.s.). For both men and women of high autonomy, in contrast, discrepancy in perceived informativeness remained a moderately important attribute (r=.61 for men, r=.48 for women). With regard to knowledge of the specific disease, the effects were somewhat different. Informativeness discrepancy was of high importance in predicting physician preference for low knowledge men (r=.76, p<.01) but of lowest importance for (r=.40, n.s.) for low knowledge men. Informativeness was of moderate importance for women, whether they had high (r=.58, p<.01) or low knowledge (r=.56, p<.01). To some degree, then, both level of autonomy and the level of knowledge affect what attributes are important.

Examination of the open-ended reasons provided by patients underlying their preferences explores this pattern further. The most frequent reasons given by those who chose the paternalistic physician was that the physician: a) showed more concern, b) simplified the situation, and c) had superior knowledge regarding treatment. The most frequent reasons provided for the choice of an informative physician was that the physician: a) gave more choices and b) provided more options.

Discussion
These results are quite rich. As is indicated in Figures 1 and 2, autonomy level and degree of knowledge of the specific disease seem to affect physician style preference for female patients but not for male patients. At the same time, autonomy and degree of knowledge of the specific disease are slightly positively related for men and significantly negatively related for women. (refer to Table 1 page 54)

This pattern suggests that autonomy may function differently for men and women in our sample. For male patients, autonomy is unrelated to fear of attachment. Acquiring knowledge from others about a disease may be predicated on an openness toward interaction with them (i.e., overcoming one's fear of attachment). Autonomous men may be more open to listen to others about their disease than nonautonomous men. Nonautonomous women, in contrast, may be open to hear about their disease than autonomous women.

More generally, this research suggests that the doctor's role cannot be understood in a vacuum, for a particular style taken by him or her may or may not be appropriate depending on the patient's personality and awareness. The effectiveness of the particular doctor's style, whether paternalistic or informative, must always be judged against the patient's autonomy level and/or knowledge.

Future research in this important area should examine the resiliency of the reported effects across different disease states, psychiatric as well as medical. Finally, gender may play a critical role here, both with regard to the patient and physician. Such sensitivity seems to be consistent with the highest ends of medical practice. -E&M

References

1 Veatch RM. A Theory of Medical Ethics. New York: Basic Books Inc. Publishers. 1981.

2 Macklin R. Moral Choices. New York: Pantheon Books Inc. 1987.

3 President's Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research. 1982.

4 Emanuel EJ, Emanuel LL. Four models of the physician-patient relationship. JAMA. 1992;267:2221-2226.

5 Richards T. Partnership with patients. BMJ. 1998;316:85-86.

6 Katz J. The Silent World of Doctor and Patient. New York: Free Press. 1984.

7 Cassileth BR, Zupkis RV, Sutton-Smith K, March, V. Information and participation preferences among cancer patients. Ann Intern Med. 1980;92(6):832-836.

8 Ende J, Kazis L, Ash A, Moskowitz MA. Measuring patients' desire for autonomy: decision making and information seeking preferences among medical patients. J. Gen Intern Med. 1989;4:24-30.

9 Kapp MB. Health Care Decision Making by the Elderly: I get by with a little help from my family. Gerontol. 1991;31:619-623.

10 Blanchard CG, Labreque MS, Ruckdeschel JC, Blanchard EB. Information and decision making preferences of hospitalized adult cancer patients. Soc Sci Med. 1988;27:1139-1145.

11 Degner LF, Sloan JF. Decision Making During Serious Illness: What role do patients really want to play? J Clin Epidemiol. 1992;45:944-950.

12 Beaver K, Luker KA, Owens R, Leinster SJ, Degner LF, Sloan JA. Treatment decision making in women newly diagnosed with breast cancer. Cancer Nurs. 1996 Feb;19(1):8-19.

13 Degner LF, Kristjanson LJ, Bowman D, Sloan JA, Carriere KC, O'Neil J, et al. (1997) Information needs and decisional preferences in women with breast cancer (The Patient-Physician Relationship). JAMA. 1997;277(18):1485-1492.

14 Davison BJ, Degner LF. Empowerment of men newly diagnosed with prostate cancer. Cancer Nurs, 1997;20(3):187-196.

15 Miller, P. and Ingram, J. G. (1976) New Measures of Daily Coping: Development and Preliminary Results. Soc. Psychiat. 11, 51-58.

16 Hirsch BJ. Psychological dimensions of social networks: a multimethod analysis. Am. J. Soc. 1979;78:812-835.

17 Staker JP, Wilson DG. The inventory of socially supported behaviors: dimensionality, prediction and gender differences. Am. J. Commun. Psychol. 1984;12:53-69.

18 Stone AA, Neal JM. New measures of daily coping: development and preliminary results. J. Per. Soc. Psychol. 1984;46:892-906.

19 Flaherty J, Richman R. (1989) Gender differences in the perception and utilization of social support: theoretical perspectives and an empirical test. Soc. Sci. Med. 1989;28(12):1221-1228.

20 Vaux A. Harrison D. Support network characteristics associated with support satisfaction and perceived support. Am. J. Commun. Psychol. 1983;13:245-268.

21 Kaplan KJ. TILT: Teaching individuals to live together. Transactional Analysis Journal. 1988;18:221-230.

22 Kaplan KJ. TILT for couples: Teaching couples to grow together. Transactional Analysis Journal. 1990;20:229-244.

23 Kaplan KJ. TILT: Teaching Individuals to Live Together. London: Taylor and Francis. 1998.

24 Kaplan KJ, Maldaver M. Parental marital pathology and completed adolescent suicide. Omega. 1993;27:131-154.

The authors would like to acknowledge the support of the Richard J. Barber Fund for Interdisciplinary Legal Research at Wayne State University and the Research Infrastructure Support Program (RISP) on Women and Gender of the Department of Psychiatry of the University of Illinois College of Medicine. We also wish to thank Mithun Nallari, Joyce Li, and Prithi Karmali, all summer fellows in the Summer Fellowship Program at Michael Reese, for their help in developing and studying the doctor-patient relation.

Kalman J Kaplan, PhD, Professor in the Department of Psychology Department of Psychology, Wayne State University and in the Department of Psychiatry, University of Illinois College of Medicine.

Martin Harrow, PhD, Professor in the Department of Psychiatry, University of Illinois College of Medicine.

Mark E Schneiderhan, PharmD, Assistant Professor in the Department of Psychiatry, University of Illinois College of Medicine and in the Department of Pharmacy Practice, University of Illinois College of Pharmacy.

Russell Omens, PsyD, Professor at the University of Illinois at Chicago and Wayne State University.

This article appeared in Volume 18:1 of Ethics & Medicine.