When Patients Lack Capacity

Abstract

Current approaches to end-of-life decision making are widely considered inadequate. We explored these complexities by examining how patients with terminal diagnoses would choose to involve their physicians and loved ones in making medical decisions, assuming they were able and unable to participate. Cross-sectional interviews of 130 patients recently diagnosed with fatal conditions were conducted. Patients were recruited from two academic medical centers using a modification of the Decision Control Preferences Scale, ranging from independent decision making to decision making that relies upon others. Patients were asked how they would balance their own wishes relative to the input of physician and loved ones in making medical decisions, and to weigh the input of loved ones relative to physician. Most patients (52%), assuming they had the capacity, would opt to share decision making with their physicians, but 15% would defer to their physicians and 34% would make decisions independently. Similarly, 44% would share decision making with their loved ones, but fewer (6%) would defer to their loved ones. Thirty-nine percent would rely upon their physicians’ judgments about what would be best for them rather than their own wishes if they became unconscious, compared with 15% who would do so if they were conscious (P < 0.001). Nonetheless, patients were more likely to weigh their loved ones’ input more heavily than their physicians’ input if they were unconscious (33%) than if they were conscious (7%, P = 0.05). Race, religion, gender, diagnosis, and health status were largely unassociated with patients’ decision control preferences. Patients with terminal diagnoses report a wide diversity of decision control preferences, but most would opt to share decision making with their physicians and loved ones. If unable to decide for themselves, they shift toward greater reliance on physician input relative to their own wishes but would weigh loved ones’ input more heavily than physician input. Deciding for patients who cannot speak for themselves may be more complex than has previously been reflected in law, policy, or clinical ethics.

Introduction

In the United States, when patients cannot speak for themselves, the standard approach of ethics, legislation, and common-law juris-prudence has been that physicians should base medical decisions upon patients’ autonomous preferences, either as stated in formal advance directives or as elicited through substituted judgments.^1-3 However, significant problems have been identified in applying this model. First, very few patients have executed formal advance directives.⁴ Second, multiple studies have shown that substituted judgments are imperfect. Surrogates frequently cannot say what incapacitated patients would want were they able to speak for themselves,^5-11 and at least one attempt to improve the accuracy of substituted judgments through formally arranged patient-surrogate communication has failed.⁵ Third, several studies have revealed great variability in individual preferences for control over how health care decisions are made with a physician.^12-15 Fourth, studies have shown that only a minority of patients would choose to have their advance directives followed strictly,¹⁶ and a slight majority of patients would opt to have the judgments of their loved ones guide decision making even if this were to contradict a hypothetically perfect living will.¹⁷

Influenced by these findings and by theoretical reflection, some ethicists now suggest that the traditional view should be modified. Some have suggested giving loved ones a more central role;^18-22 others have emphasized the dialogical nature of decision making;^23,24 still others have recommended following a more inclusive principle than autonomy, such as respect for persons or respect for integrity.^25,26

The present study was undertaken to help understand the complexity of decision making at the end of life by studying the decision control preferences of patients with terminal diagnoses. Decision control preferences do not reflect what medical interventions patients would choose, but how they would choose to have these decisions made, ranging from an independent style to a style reliant upon the judgments of others about what would be best for them.^12,27 Our study was designed to add to the literature in three ways. First, in contrast to most previous studies, we examined the decision control preferences of patients recently diagnosed with conditions that are generally fatal, for whom these issues had authentic urgency and importance. Second, we endeavored to provide a more comprehensive picture by simultaneously examining the role patients would view as most appropriate for all three of the parties most often involved in decision making—the patient, the patient’s loved ones, and the patient’s physician. Third, we looked simultaneously at patients’ decision control preferences, both assuming that they could participate and assuming that they were unconscious and could not participate in the decision-making process, comparing both perspectives.

Methods

Sample

The study institutions were The Johns Hopkins Medical Institutions in Baltimore, Maryland, and the Saint Vincent Catholic Medical Centers, St. Vincent’s Hospital-Manhattan in New York City. Patients who met the study criteria and were being treated as either inpatients or outpatients by medical specialists at one of the two study institutions were referred to the investigators between April 2000 and June 2003. Inclusion criteria were as follows: at least 18 years of age, English- or Spanish-speaking, and interviewed within 8 weeks of being diagnosed with amyotrophic lateral sclerosis (ALS) or advanced cancer, or within 8 weeks of being hospitalized for class III-IV congestive heart failure (CHF). Cancer patients included those with nonresectable non-small-cell lung cancer, stage III or IV pancreatic cancer, or stage IV liver, colon, or gastric cancer. CHF patients must not have been eligible for transplantation. These criteria were selected to obtain a sample of patients with a prognosis of at least 50% 2-year mortality based on the literature^28-31 and the clinical experience of the respective medical specialists at the two study sites. Patients were screened and excluded for altered mental status. The two study sites were selected to obtain a sample that was racially, ethnically, and culturally diverse.

Instruments

To determine how patients would choose to have medical decisions made regarding their care, we used a modified version of the Control Preferences Scale, originally developed and validated by Degner and coworkers.^12,27 This instrument, which has now been widely used,^12-15,32 employs picture cards with explanatory text that depicts a five-point scale of decision control ranging from independent (A, B) through shared decision making (C) to decision making that is reliant on the physician or loved ones (D, E). Using these sets of cards, patients indicated their preferred level of decision control relative to the physician and relative to the loved ones. Degner’s original cards were developed only to elicit patients’ decision control preferences under the assumption that the patients had decision-making capacity. Our modification, developed with the assistance of Degner, created cards that would reflect decision control preferences under the assumption that the patients were unconscious and unable to speak for themselves. The text and cartoons depict the act of ascertaining the patient’s wishes in the unconscious state as an act of substituted judgment, contrasted with deciding in the patient’s best interests. We chose unconsciousness as a diagnosis for these scenarios so that it would be clear that the patient lacked all decision-making capacity (see Appendices I and II).

We also asked patients to weigh the input they would choose to assign to the physician relative to loved ones, both under the assumption that they were able to speak for themselves and under the assumption that they were unconscious (see Appendix III).

In addition, we administered four other previously validated instruments: 1) the McGill Quality of Life Questionnaire,^33-36 2) the General Health Questionnaire (GHQ),^37-39 3) the Short Portable Mental Status Questionnaire (SPMSQ),^40-43 and 4) the Confusion Assessment Method (CAM), an observational rating system for assessment of delirium.⁴⁴ We used validated Spanish translations of instruments where available and developed our own translations where necessary.⁴⁵ All instruments were administered by trained research assistants, one of whom was fluently bilingual.

Procedures

Of the 288 patients found eligible for the study, 130 (45%) gave usable interview data. Of the 158 who did not give usable interview data, 108 (68%) refused; 33 (21%) could not be reached; 5 (3%) died before they could be interviewed; 5 (3%) had family members who controlled access to the patient by phone or physical presence and refused to allow us to ask the patient for consent to participate; 3 (2%) were unable to complete the interview; 2 (1%) did not participate for other reasons; and 2 (1%) were excluded because they had missing data for the Control Preferences questions. There were no significant differences between participants and nonparticipants in race (P = 0.66), sex (P = 0.25), diagnosis (P = 0.34), age (P = 0.56), or study site (P = 0.27). Participants were assured of confidentiality, with the exception that if they indicated they would harm themselves, the investigator would inform their attending physician. All consenting patients had the cognitive capacity to participate as determined by their scores on the SPMSQ and the CAM. The study was approved by the Johns Hopkins Medicine Institutional Review Board and by the St. Vincent’s Hospital-Manhattan Institutional Review Board.

Data Analysis

We used descriptive statistics to summarize the demographic, quality-of-life, and general health survey data. Taking the subjects’ highest ranking choice of decision control (A, B, C, D, or E), we used χ², Fisher’s exact, and Spearman ρ testing to assess differences among patient responses to the four scenarios according to various clinical and sociodemographic variables. To compare patient responses in different scenarios, we trichotomized the data, combining the two independent responses (A, B) and the two reliant responses (D, E) while retaining the shared decision-making response (C). Patients’ decision control preferences (both with respect to physician and with respect to loved ones) and their weighting of the relative input of physician and loved ones regarding decisions made under the assumption they were able to speak for themselves were compared with their responses under the assumption they were unconscious using the Stuart-Maxwell test—an extension of the McNemar test for greater than two categories.⁴⁶ Our sample size of 130 subjects gave us a power greater than 99% to detect a shift in preferences from the conscious to the unconscious state at α = 0.05.⁴⁷ All calculations were performed using the SPSS software package except the Stewart-Maxwell test, which was programmed into Excel by our statistician (RT).

Results

Characteristics of the Sample

As Table 1 shows, the sample was racially diverse, with 64% white, 23% black, and 12% Hispanic or other. The majority of the study participants were male (63%) and did not have any formal education beyond high school (65%). Most patients (90%) identified themselves as Christian (Catholic, Protestant, Orthodox, or other Christian), with other patients identifying themselves as either Jewish or as having no religion. No patient identified him/herself as Muslim or as belonging to another religion. About one half of the patients were married, and the mean age of study participants was 62 years. Most participants (66%) received Medicare or had private insurance, and 39% reported they had completed an advance directive. Forty percent had CHF, 36% cancer, and 24% ALS. Participants’ subscale and overall mean scores for the GHQ and the McGill questionnaire are reported in Table 2.

Table 1. Patient Characteristics.

Variable (Sample Size)	Percent (Frequency) or Mean (SD)
Site (n = 130)
Johns Hopkins	54.6% (71)
Saint Vincent’s	45.4% (59)
Mean age in years (n = 129)	62.0 (12.5)
Sex (n = 128)
Male	63.3% (81)
Female	36.7% (47)
Race (n = 129)
White	64.3% (83)
Black	23.3% (30)
Hispanic	7.0% (9)
Other	5.4% (7)
Diagnosis (n = 130)
CHF	40.0% (52)
Cancer	36.2% (47)
ALS	23.8% (31)
Religion (n = 128)
Christian	89.8% (115)
Jewish/other/none	10.2% (13)
Frequency of religious services (n = 130)
Once a month or more	54.6% (71)
Less than once a month	45.4% (59)
Significance of religion (n = 130)
Not important	11.5% (15)
Somewhat important	17.7% (23)
Very important	70.8% (92)
Education (n = 130)
High school or less	64.6% (84)
College or greater	35.4% (46)
Marital status (n = 130)
Married	49.2% (64)
Not married	50.8% (66)
Advance directive (n = 130)
Yes	38.5% (50)
No	61.5% (80)
DNR order of record (n = 127)
No	96.9% (123)
Yes	3.1% (4)
Insurance (n = 129)
Medicaid/none	24.0% (31)
Medicare/private	65.9% (85)
Other	10.1% (13)

DNR = Do not resucitate.

Table 2. GHQ and McGill Scores^a.

Subscale	Mean	SD
GHQ
Somatic (n = 128)	7.5	4.4
Anxiety (n = 130)	7.1	5.6
Social dysfunction (n = 128)	9.8	3.8
Depression (n = 130)	3.2	4.5
Total score (n = 127)	27.5	15.0
McGill
Physical symptoms (n = 130)	5.8	2.8
Physical well-being (n = 130)	6.4	2.5
Psychological (n = 130)	7.2	2.6
Existential (n = 130)	7.8	1.6
Social support (n = 130)	8.6	1.8
Total score (n = 130)	7.2	1.6

^aHigher scores on the GHQ indicate greater levels of distress. Higher scores on the McGill questionnaire indicate higher quality of life.

Decision Control Preferences

Patients’ decision control preferences, both with respect to the physician and with respect to loved ones, both under the assumption that they were able to participate and under the assumption that they would be unconscious, are summarized in Figs. 1 and 2.

Fig. 1.

Patient-physician dimension: conscious vs. unconscious.

Fig. 2.

Patient-loved ones dimension: conscious vs. unconscious.

When Able to Participate

When able to speak for themselves, the majority of these patients with terminal diagnoses (52%) would make decisions in collaboration with the physician (choice C), while 34% of patients would decide independently of the physician (choice A or B) and 15% would delegate decision-making responsibility to the physician (choice D or E). With respect to loved ones, half of the patients (50%) would make decisions independent of their loved ones, 44% would make shared decisions, and only 6% would rely upon their loved ones.

When asked (assuming they themselves were able to participate in the decision-making process) whose opinion should weigh more, the physician’s or their loved ones’, 51% of patients chose both equally, 42% the physician’s, and only 7% their loved ones’.

When Unable to Participate

When asked (assuming that they should become unconscious) how much weight the physician’s input should have relative to their own wishes for treatment (as previously stated or otherwise understood), a significant shift was noted when compared with their choices assuming they could speak for themselves (Fig. 1). The number of patients who would rely upon their physicians’ judgments in making decisions assuming that they were to become unconscious (39% choosing D or E) was more than the number of those who would rely upon their physicians’ judgments assuming that they themselves were capable of participating in the decision (15% choosing D or E, P < 0.001).

When asked (assuming they should become unconscious) how much weight loved ones’ input should have relative to their own wishes for treatment, a complex pattern of changes was observed when comparing the conscious with the unconscious scenarios. Assuming they were to become unconscious, more patients would rely upon their loved ones’ judgments (21% choosing D or E) than if they were capable of participating in the decision themselves (6% choosing D or E, P = 0.39). Moreover, a careful look at Figure 2 shows that while the overall proportion who favored independence from loved ones (choice A or B) was the same in the unconscious scenario (63/130) as in the conscious scenario (65/130), the proportion expressing more complete independence (choice A) was higher in the unconscious scenario (29/130) than in the conscious scenario (14/130).

When asked how they would weigh the relative input of the physician vs. loved ones on the assumption that they were unconscious and unable to participate in the decision, 48% would weigh both equally, 33% would weigh their loved ones’ input more heavily, and 19% would weigh the physician’s input more heavily (Fig. 3). Compared with the conscious scenario, this represents a shift toward weighing the input of loved ones more and physician input less (P = 0.05). Totally, 55% changed their choice from the conscious to the unconscious scenario. The vast majority (88%) of those who changed their choices moved in a direction of increasing involvement of loved ones, and 10 patients switched completely from weighing the physician’s input most heavily in the conscious scenario to weighing loved ones’ input most heavily in the unconscious scenario.

Fig. 3.

Physician-loved ones dimension: conscious vs. unconscious.

Factors Associated with Control Preferences

Of the demographic variables listed in Table 1, there was a single nonsignificant trend (P = 0.06) for association with decision control preferences: patients who identified themselves as some form of Christian were somewhat more likely to value the opinion of loved ones compared with those who identified themselves as Jewish or belonging to no organized religion. Of the clinical factors listed in Table 1 and the health scales in Table 2, only diagnosis had any association with control preferences. When considering the conscious state, patients with cancer were more likely to choose a decision-making style that was independent of their loved ones compared with patients who had ALS or CHF (64% in cancer vs. 45% in ALS vs. 40% in CHF, P = 0.01). There was also a nonsignificant trend for patients with cancer to be more likely to choose a decision-making style that was independent of their physicians (43% in cancer vs. 26% in ALS vs. 31% in CHF, P = 0.07). Diagnosis was unassociated with control preferences in the unconscious state. Quality of life, as measured by the McGill questionnaire, and functional health status, as measured by the GHQ, were also unassociated with decision control preferences.

Diversity of Control Preferences

We looked at the frequencies of how individual patients’ decision control preferences were distributed along all three control dimensions—that is, how individual patients answered the patient-physician, patient-loved ones, and physician-loved ones questions. We found remarkable diversity. Considering the patient-loved ones, patient-physician, and physician-loved ones questions, there were 75 possible patterns of response for any individual. Patients’ decision control preferences were distributed across 43 of these 75 possible combinations assuming they could participate in decision making, and across 53 of these 75 possible combinations assuming they were unable to participate.

Discussion

Most studies of decision making at the end of life have explored what treatments should be given to patients. This study used a relatively simple test of how patients with generally fatal diagnoses would opt to have their health care decisions made. The results reveal just how complex this process can be. While a plurality opted for shared decision making in all scenarios, substantial minorities opted for either independent decision making or decision making that would rely upon others’ judgments. Individual patients varied in complex ways. For example, many patients, assuming they retained capacity, opted for decision making that would be independent with respect to their loved ones but reliant with respect to their physicians. Age, gender, race, and religion did not predict how patients would opt to have decisions made. Should they become unable to speak for themselves, patients were more likely to opt for greater reliance upon physicians, were divided about the role of their loved ones, but would weigh the input of their loved ones more heavily than that of their physicians.

Conventional legal and ethical theory counsels physicians to provide the loved ones of all incapacitated patients with nondirective information and to ask loved ones to direct the decision by exercising a pure substituted judgment based upon the patient’s autonomous wishes.⁴⁸ Our results suggest that a physician who did this uniformly for all patients would be employing a style of decision making that would diverge from the styles most patients would claim as their own.

Physicians’ and Loved Ones’ Roles When Patients Can Speak for Themselves

As in studies of patients newly diagnosed with curable malignancies,^12-15,32 we found that most of our patients with fatal diagnoses would make medical decisions in collaboration with the physician when conscious, but a significant minority would decide independently and a smaller minority would delegate decision making to the physician.

We found that when considering the role of loved ones, patients were almost equally divided between those who would make decisions in collaboration with loved ones and those who would be independent of loved ones. Relatively few patients (6%) would rely largely on loved ones.

When asked how they would balance the input of the physician relative to that of loved ones, more patients would give greater weight to the physician’s opinions than their loved ones’ opinions. However, half the patients (51%) would weigh both the doctor’s and loved ones’ input equally.

These findings partially support, but also partly undermine, claims that the family has a large role in making medical decisions for patients with capacity in the United States.^18,19 A plurality of patients preferred a balance of their own independent preferences and the wishes of their loved ones. However, very few patients, if able to speak for themselves, would defer largely to loved ones, and a substantial minority would assign greater importance to the input of the physician than that of their loved ones.

Physicians’ and Loved Ones’ Roles When Patients Cannot Speak for Themselves

The roles patients would assign to physician and loved ones shifted significantly, however, under the assumption that they were unconscious. In the unconscious scenarios, Choices A and B (independent) indicate that the patient is opting for substituted judgment, while Choices D and E (reliant) indicate that the patient is opting for a best interests standard. Comparing patients’ responses should they lose decision-making capacity with their responses should they retain capacity (Fig. 1), we found that patients shifted toward giving greater weight to the physician’s judgment relative to their own preferences. This suggests that while patients tend to opt for shared decision making with the physician when they are capable of making decisions, many see the physician’s role, should they lose this capacity, as making decisions based upon their best interests rather than attempting substituted judgment.

In the absence of good qualitative data, it is difficult to interpret why many patients (should they become unable to participate in decision making themselves) would choose to have their physicians make best interests judgments. This may reflect patients’ low confidence in their physicians’ ability to know them well enough to make substituted judgments, and so they would prefer that physicians “stick to the medical facts” and make medical best interests judgments. More likely, however, this may reflect a faith in physicians’ knowledge and clinical experience. Thus, patients may believe that physicians’ proper role is to provide good advice and information to help their loved ones make good decisions.

A plurality of patients, whether conscious or unconscious, opted for a decision-making style that would be independent of their loved ones. Nonetheless, should they lose capacity, 31% would opt for a mixture of best interests and substituted judgment, and a substantial minority (21%) would lean toward having medical decisions based upon their loved ones’ judgments about their best interests.

These findings differ from those of Degner and Sloan,¹² who found that only 10% of Canadians would choose to have their families make decisions based on a best interests standard rather than a substituted judgment standard should they become unable to speak for themselves. Whether this reflects differences in culture between the United States and Canada, the fact that we interviewed patients who had terminal diagnoses rather than persons who were well, or other factors is uncertain.

In addition, it is known that even some patients who would opt, should they lose capacity, for having their surrogates perform a substituted judgment may be doing so not to express their autonomy but as a means of relieving their surrogates of the burden of making a “best interests decision.”^17,49 These considerations suggest that a process far more complex than a desire for autonomous preference satisfaction is at work in end-of-life decision making.

Sociodemographic and Clinical Variables

We found few statistically significant associations between clinical and sociodemographic variables and decision control preferences. This contrasts with the findings of other studies.^12,15,50 However, even in these studies, sociodemographic variables accounted for less than 15% of the total variance.^12,27 Our data suggest that there are as many differences within ethnic and religious groups as there are between them, and that patients’ capacity to participate in decision making is the over-whelming predictor of the patient’s decision control preferences. This may serve as a reminder against making assumptions about a patient’s decision control preferences based on demographic characteristics such as age, sex, education, religion, or ethnicity.

Diversity of Patient Decision Control Preferences

One of the most striking findings of this study is the sheer diversity of patient choices across these three control dimensions. Although the shared decision-making role was most popular—along each dimension and in each scenario, it was always selected by more than 30% of the participants—few patients picked shared decision making in all three scenarios. Altogether, the 130 study participants fell into 43 and 53 of the 75 possible combinations in the conscious and unconscious scenarios, respectively. This suggests that when patients’ decision control preferences are expanded to consider these three control dimensions simultaneously rather than considering only the single patient-physician dimension (the dimension that has been investigated most extensively), the diversity of control preferences expands dramatically.

Limitations

The study was conducted at only two hospitals and so may not be generalizable to other settings. However, the two sites were chosen to obtain a sample that is ethnically, culturally, religiously, and economically diverse. While the patients we studied may not be representative of all patients with terminal conditions, we chose diagnostic categories with different disease trajectories (i.e., CHF, ALS, cancer) so that they might reflect patients’ experiences with a broad range of diseases.

One might also question the validity of the instruments used in the study. However, the Decision Control Preferences Scale was validly derived from the grounded theory on which its construct is based.²⁷ Additionally, it has now been widely used in empirical studies of patient decision making.^{12,27,32,49,51,52}

Implications

Our findings suggest that health professionals who care for patients with terminal diagnoses should not assume that the weight patients place upon their own autonomous preferences when they can participate in decision making is the same as the weight they would choose to have accorded to their previously stated wishes should they become too ill to speak for themselves. Discerning the composite weight each patient would give to his or her own preferences, his or her loved ones’ input, and physician input is a complex process. When patients lack capacity, this information can usually only be gathered by careful conversations with families and other loved ones. Advance care planning might help this process if it were to attend not just to what patients would want but to how they would want decisions to be made. Nurses, who tend to spend more time with patients and their loved ones, can play an important role in this process. Above all, physicians should be aware that the paradigm pattern of providing nondirective information to loved ones who then make the medical decision by exercising substituted judgment is simply incongruous with the decision-making style of most patients.

It is important to note that these data do not settle any normative questions about bioethics or U.S. law. Choosing to rely on others for decision making may not be incompatible with the autonomy model—a patient may meta-autonomously choose to rely on others for decision making. However, the dominance of the autonomy model in legal and clinical practice may present obstacles to patients whose decision-making style is not autonomy centered. Many ethicists think that it is normatively misguided simply to equate the enhancement of patient autonomy with morally better decision making.^18-22,25,26 Law, policy, and ethics ought to accommodate not only patients who would have their autonomous preferences direct their care but also those who would defer extensively to loved ones and physicians in making medical decisions.⁵³ Policy and practice might better serve patients if they were to reflect the enormous complexity that instruments, such as advance directives, can only inadequately express.

Appendix I

Control Preferences Scale’s Text (Adapted from Degner and Sloan12)

Control Preference		Patient-Physician/[Patient-Loved Ones] (Conscious)	Patient-Physician/[Patient-Loved Ones] (Unconscious)
Independent	A	“I prefer to make the decisions about which tests or treatments I receive.”	“I prefer that my doctor order tests or treatmentsfor me [loved ones tell my doctor which tests or treatments to order] based on my own personal wishes.”
	B	“I prefer to make the final decision about which tests or treatments I receive after seriously considering my doctor’s opinion [loved one’s opinions].”	“I prefer that my doctor order tests or treatments [loved ones tell my doctor which tests or treatments to order] for me based on my own personal wishes after having seriously considered what he or she thinks [they think] is best.”
Shared	C	“I prefer that my doctor [loved ones] and I share responsibility for deciding which tests or treatments I receive.”	“I prefer that my doctor decide about [loved ones tell my doctor] which tests or treatments to order by equally weighing both my personal wishes and what my doctor [loved ones] thinks is best.”
Reliant	D	“I prefer that my doctor [loved ones] make the final decision about which tests or treatments I receive after seriously considering my opinion.”	“I prefer that my doctor [loved ones] order tests or treatments for me based on what my doctor [loved ones] thinks is best, after having seriously considered my personal wishes.”
	E	“I prefer to leave all decisions about which tests or treatments I receive to my doctor [loved ones].”	“I prefer that my doctor order tests or treatments [loved ones tell my doctor which tests or treatments to order] for me based on what he or she thinks [they think] is best.”

Appendix II

Appendix III

Physician vs. Loved Ones Questions

In making decisions about your health care now, how do you weigh the input of your doctor and the input of your loved ones?

1 = My doctor’s input weighs most heavily.

2 = My loved ones’ input weighs most heavily.

3 = My doctor’s input and my loved ones’ input are about equally important.

If you were to become too sick to be able to speak for yourself, whose input should weigh most heavily in making decisions about your health care?

1 = My doctor’s input weighs most heavily.

2 = My loved ones’ input weighs most heavily.

3 = My doctor’s input and my loved ones’ input are about equally important.