The importance of patient preferences in treatment decisions—challenges for doctors

Introduction

Health professionals are increasingly encouraged to involve patients in treatment decisions, recognising patients as experts with a unique knowledge of their own health and their preferences for treatments, health states, and outcomes.1 2 Increased patient involvement, a result of various sociopolitical changes,^w1 is an important part of quality improvement since it has been associated with improved health outcomes3 ^w1-w9 and enables doctors to be more accountable to the public.

However, this poses challenges for doctors. We discuss these in relation to the competences for shared decision making that have been proposed.4 ^w10

Methods

We made literature searches using Medline, Web of Science, PsychINFO, CINAHL, the Cochrane Library, and HMIC (key words “consumer participation,” “patient participation,” “decision making,” “patient preferences,” “shared decision making,” “patient involvement in decision making”). We also searched references of articles, indexes of key journals, important texts about patient involvement, and key reviews.

We conducted informal interviews with doctors from a range of specialties (general practice, orthopaedics, stroke medicine, accident and emergency, and vascular surgery) and recorded their opinions to provide a focus to this discussion (quotes in italics).

Establishing a partnership

For patients' views about treatment options to be valued and necessary, there must be a partnership between doctor and patient, but establishing one requires both time and certain skills.

“There's not enough time”—The pressure of time is a perpetual challenge; doctors are particularly concerned about the implications of informing patients without allowing extra time for this.5 However, involving patients more in treatment decisions may have no significant effect on consultation length3: adequate discussion at an early stage may allow more succinct discussion later and ultimately save time.6 Nevertheless, it may be difficult to overcome doctors' perceptions of the extra time required.

Doctors may not have the appropriate communication skills to elicit patients' preferences and involve them in treatment decisions^w10; for example, general practice registrars would welcome further training in involving patients in decision making.7

“You need to build a relationship up over time”—Doctors may find it easier to share decision making with patients they know well.8 ^w11 However, a patient may not see the same doctor at every consultation, and care may be multidisciplinary. Doctors seeing a patient for the first, or a single, time may have to be particularly skilful at gaining the patient's trust and eliciting his or her preferences, which are likely to be affected by the patient's experience of other doctors.9

Eliciting patients' preferences

“It's difficult to generalise about what role patients prefer in treatment decisions”—Doctors may have a poor understanding of their patients' preferences for involvement in decision making.10 In one study doctors wrongly estimated hypertensive patients' preferences for discussion about treatment 40% of the time, underestimating this and their desire for information, while overestimating their desire to participate in decisions.11 Finding ways to elicit patients' preferences is therefore a considerable challenge. Furthermore, preferences may be influenced by the way that they are elicited,12 and doctors may elicit preferences from certain groups of patients more readily than others.

Although some patients may not wish to make the final choice of treatment, many would prefer more information.^w12-w14 Deber suggested there may be two components of treatment decisions—problem solving (“identifying the one right answer”) and decision making (“selecting the most desired bundle of outcomes”)—and hypothesised that, whereas patients may prefer doctors to perform the problem solving component (which requires clinical expertise), patients would want to be involved in decision making.13 This was supported in a survey of patients undergoing angiography.^w15

“Some patients just don't want to be involved”—Some patients may not want to participate in decision making.^w16 This may be because they feel they lack the knowledge and experience to develop informed preferences and are afraid of making the wrong decision.^{w17 w18} Nevertheless, it is thought that most patients want doctors to understand their preferences even if they do not wish to make the final decision.14 Furthermore, patients' role preferences may change over time and through the course of an illness,15 so doctors have to be sensitive to such changes.

Giving patients information

Patients must be given technical information that is clear and unbiased to ensure that their preferences are based on fact and not misconception.

“People don't retain information”—Even when patients are given such information they may not recall it. In a study of patients' recall of the risks associated with endarterectomy, patients recalled risks ranging from 0 to 65% for a communicated risk of 2%.16

“You have to be able to give the right information at the right time”—Doctors feel that a lack of suitable information, or of access to it, is a barrier to shared decision making,7 while there seems to be too little information designed specifically to support patient involvement.17

Presenting risk

“You can always present information so they select the treatment you want them to”—A study of nephrologists revealed that doctors used information to influence patients' treatment choices.18 Presenting “logically equivalent” information in different ways has a substantial effect on treatment preferences.19 ^w19 For example, relative risk is more persuasive than absolute risk.6

Enabling patients to understand risks is crucial before considering different treatment options. Yet risk is a complex phenomenon that many patients (and doctors) find difficult to understand. Common errors include compression bias (the tendency to overestimate small risks and underestimate large ones), miscalibration bias (overestimation of the level and accuracy of one's knowledge), availability bias (overestimation of notorious risks20), and optimism-pessimism bias (the tendency of patients to believe that they are at less risk of an adverse outcome than people similar to them14).

Patients may understand different methods of presenting risk to varying extents. Numerical information is often poorly understood,6 ^w20 so doctors may need to determine how to present risk in a way that is helpful for an individual patient, or at least be able to present it in different ways. Differences in education and experience may also affect patients' understanding of risk; methods of increasing comprehension among people with poor literacy and numeracy will be essential.21

Individualised risk information—that is, based on each patient's clinical characteristics—is available for only a few conditions, such as the Framingham equation for stroke or cardiovascular disease.^w21 Even this information is probabilistic in nature and requires skilful communication.

Furthermore, there may be ambiguity associated with the language of risk. Doctors and patient may have different perceptions of what is “low risk,” “unlikely,” or “rare.”6 ^w22 There may also be variation in acceptability of risk, so doctors need to ascertain the degree of risk that a patient is comfortable with to ensure discussion is relevant. In addition, people's conceptions of a condition and its implications may vary.6

Negotiating a decision in partnership with the patient

In many cases a shared decision that is acceptable for both patient and doctor will arise from discussion, but sometimes this is more problematic.

“It's hard to integrate patients' preferences if they have unreasonable expectations”—Reaching consensus is difficult if patients' preferences do not match those of doctors. Differences between patients and doctors have been shown in many clinical settings, including cardiovascular disease, cancer, obstetrics and gynaecology, and acute respiratory illness.23 Patients' preferences may also disagree with clinical guidelines,^w25 which may have implications for concordance and poses a challenge to doctors to resolve such discrepancies.

Some doctors are concerned that patients will demand treatments that offer little benefit but which may be expensive. The conflict between individual and societal needs is particularly important in the publicly funded NHS,^w26 and strategies will have to be developed to deal with inequitable demands.

Inequalities in patient choice

A further concern is that residual racial and sex discrimination, along with socioeconomic and educational inequality, may cause groups of patients to have systematic variation in their preferences and in their capacity to articulate them.24 Black patients with HIV were less likely to have discussed their treatment preferences with their doctor than white HIV positive patients.25 Access to services is affected by sex, race, and socioeconomic status,24 which may affect expectations and perceptions of risk. In women with menstrual disorders, educated women were much less likely to agree to hysterectomy.10 This may produce further widening of inequalities, particularly if certain groups of patients are more vocal about their preferences and demanding of particular treatments.

Conclusions

To improve the quality of care they provide, doctors should understand their patients' preferences. However, this raises many challenges for doctors. Practical concerns include time pressures and difficulties in eliciting preferences from patients who may be hesitant to make treatment decisions. These are compounded by a deficit of appropriate information to support patients' decisions. Doctors may not have the appropriate interpersonal skills, particularly for communicating risk. Medical uncertainty, deficiencies in individual doctors' knowledge, and the highly variable ability of patients to understand and remember clinical information mean that risk communication is often inadequate to support patients in making informed decisions.

Some doctors may wish to uphold the imbalance of power between themselves and their patients, which may make patients reluctant to share their preferences. Inequalities in health care may be perpetuated or exacerbated if these affect patients' preferences or the extent to which doctors seek to inform or understand them. There are also challenges to be overcome if patients express preferences that contradict those of their doctors or guidelines.

Despite these challenges, many doctors are committed to understanding patients' preferences. Innovative research is being undertaken to find solutions to these problems. The imaginative use of different sources of information, together with evidence based decision aids and decision analysis, are likely to be useful.

Extra references (w1-w26) are listed on bmj.com