Review For reprint orders, please contact
[email protected]
Patient preferences and healthcare outcomes: an ecological perspective Expert Rev. Pharmacoecon. Outcomes Res. 12(2), 167–180 (2012)
Richard L Street Jr*1,2, Glyn Elwyn3 and Ronald M Epstein4 Department of Communication, Texas A&M University, College Station, TX 77843-74234, USA 2 The Houston Center for Quality of Care & Utilization Studies, Section of Health, TX 77843-74234, USA 3 Department of Primary Care & Public Health, Cardiff University, Cardiff, UK 4 Departments of Family Medicine, Psychiatry, Oncology & School of Nursing, University of Rochester Medical Center, Rochester, NY, USA *Author for correspondence: Tel.: +1 979 845 0209 Fax: +1 979 845 6594
[email protected] 1
This article examines the nature of patients’ preferences for healthcare and whether clinician accommodation of patient preferences influences health outcomes. First, we provide a conceptualization of patient preferences along with their key attributes. Second, we review research on the relationship between health outcomes and patient preferences for treatments and for the process of care (e.g., preferred involvement in decision-making). Third, following a critique of this literature, we present an ecological model of patient preferences that, while acknowledging that patient preferences may emerge from various contexts (e.g., family or media exposure), we focus on the important role that clinical encounters and patients’ health-related experiences play in the elicitation and construction of patient preferences. Fourth, we propose two pathways, one behavioral (adherence) and the other psychological (sense of autonomy or satisfaction with decision), through which meeting patient preferences could lead to better health outcomes. Fifth, we discuss how preferences can be elicited and clarified through patient-centered conversations. We conclude with implications for future research and clinical practice. Keywords : health outcomes • patient-centered care • patient preferences • physician–patient communication • shared decision-making
A fundamental tenet of patient-centered care is that treatment decisions should take into account the patient’s preferences, values and wishes [1] . Accommodating patient preferences is an extension of the ethical principles of respect for person and autonomy, which are generally regarded as ‘the right thing to do’ when providing medical care. However, researchers and clinicians are also interested in the question of whether satisfying patient preferences for medical care has an impact on health outcomes and, if so, which ones (biomedical, pain, quality of life or emotional wellbeing)? This question assumes patient preferences exist, are identifiable and can be acted upon in meaningful ways. This article reviews the evidence surrounding these issues and proposes a theoretical framework that conceptualizes patient preferences and describes pathways through which clinician–patient communication can elicit, explore and construct patient preferences in ways that could lead to better health outcomes. To accomplish this, we first provide a conceptualization of patient preferences including their various attributes. Second, we review previous research addressing the relationship www.expert-reviews.com
10.1586/ERP.12.3
between health-related outcomes and whether patients’ preferences for healthcare were enacted or incorporated in medical decisions. Third, we critique this research and present a conceptual model describing the role of clinician–patient communication in eliciting and constructing patient preferences, and the pathways through which these conversations could help clinicians and patients achieve a clearer understanding of what patients want from their healthcare, which when acted upon, can contribute to improvements in patients’ health and wellbeing. Finally, we discuss the implications of our model for research and clinical practice. Conceptualizing patient preferences
Although there has been a large body of research examining patients’ preferences for healthcare, rarely do investigators define what preferences are and their constituent attributes. As Brennan and Strombam note, although patient preferences lack a consistent definition, there is convergence in the notion that preferences are “statements made by individuals regarding the relative desirability of a range of health experiences, treatment options and health states” [2] . Put even more
© 2012 Expert Reviews Ltd
ISSN 1473-7167
167
Review
Street, Elwyn & Epstein
simply, preferences are what patients want from their healthcare. It is the task of clinicians and patients to make preferences explicit and address them in the consultation if they are germane to the quality of care the patient receives. For our purposes, what is more important than a definition of patient preferences per se is an understanding of their various dimensions. Specifically, patient preferences vary widely with respect to domain, scope, mutability and clarity. With respect to domain, investigations have studied patients’ preferences for specific treatments [3–6] , for how involved patients want to be when deciding treatment [7–12] , for the kind and amount of information patients want from their doctors [13–17] and for a physician’s communicative style [18–21] . Preferences also vary in scope. Some measures of patient preferences tap broadly into patients’ general beliefs and values about healthcare, such as a patient’s orientation to the physician–patient relationship and medical care (e.g., ‘a treatment plan cannot succeed if it is in conflict with a patient’s lifestyle or values,’ or ‘the patient must always be aware that the doctor is in charge’) [22] or their predispositions for involvement in medical decision-making (e.g., autonomist or delegator) [23] . Other measures tap into specific situations or events, such as a preference for treatment (e.g., “would you prefer medication, talking therapy, or have no preference for treating your depression?” [24]) or a particular healthcare provider or insurance plan [2] . Preferences also vary in their modifiability, depending on the patient and the health condition. For example, in some situations, preferences may be quite stable. This may be especially true for mundane, simple and familiar medical conditions [25] . However, even for serious, life-threatening conditions (e.g., cancer), patients will vary in how strongly they hold on to what they want from their healthcare. For example, in a study of women who were carriers of BRCA gene mutations, which indicate a very high likelihood of developing cancer, most patients had strong and, over time, stable preferences for a particular monitoring or management strategy [26] . However, other similarly affected women may change their minds, such that an early inclination toward vigilant screening might shift to a preference for prophylactic mastectomy after conversing with others (including clinicians), pondering family history, receiving new information or experiencing emotional stress [27] . Similarly, a patient who typically wants to be highly involved in the clinical decision-making may later want to defer decisions to others as his or her condition progresses or becomes more serious [28] . Finally, for many situations that are new or complex, such as whether to undergo another course of chemotherapy for cancer that has progressed despite treatment, or choosing between early versus late initiation of HIV medications, patients may not have clear preferences. Sometimes these situations are both unfamiliar and emotionally wrenching; patients may be conflicted in what they want from their care in ways not be easily translated into a clear preference (e.g., ‘do everything possible’ vs ‘make my life more comfortable,’ or ‘maximize effectiveness’ vs ‘minimize side effects’) [25] . In these situations, a patient’s preferences may be either poorly formulated or nonexistent and may need to be 168
informed and co-constructed through conversations with clinicians, family and others. Through conversations and deliberation, the patient can integrate his or her understanding and feelings [29] into preferences for care that can be articulated and enacted [30] . In summary, as a theoretical construct, patients’ preferences for healthcare are quite complex and range from the very clear, specific and intractable to the elusive, vague and unstable. Researchers and clinicians must understand that preferences are not only elements of a patient’s motivations and goals, they are also inextricably tied to a patient’s understanding of their health and wellbeing, and the impact of these on others. While research in cognitive psychology suggests that preferences are more likely to be unstable in situations that are novel, unfamiliar and affectively charged [31] , these assertions have not been widely tested in real clinical environments. In this sense, a patient’s preferences should be studied as highly contextualized within specific situations, relationships and problems to be solved. These issues are addressed in more detail later in this article. Patient preferences & health outcomes: a review of the state of the science
Research on patients’ preferences and health-related outcomes typically fall within two categories: those associated with treatment and outcomes; and those associated with the process of care (e.g., preferences for amount and kind of information, involvement in decision-making and the clinician’s style of communication). In this section, we review research, including summaries of previous reviews and meta-analyses that examine whether meeting patients’ preferences for care leads to better health outcomes. As a précis, it is important to note that many of the studies reviewed, or that are included in systematic reviews, are observational studies, present conflicting evidence or report relatively small effect sizes. Preferences for treatment & outcomes
Patient preferences for treatment and outcomes are intertwined, in that a patient’s desire for a particular treatment is often based on a belief or hope that more favorable outcomes are associated with that treatment [32] . For example, breast cancer patients more worried about recurrence are more likely to want mastectomy over breast conservation than are patients who are more concerned about disfigurement or lymphedema – even when they understand that life expectancy is comparable between the two [33] . Men with benign prostate disease who report that their symptoms are severe and who are less concerned about sexual dysfunction are more likely to prefer prostatectomy over ‘watchful waiting’ for treatment [34] . The question of interest in this article is whether patients who received their preferred treatment subsequently experienced better health outcomes. The evidence is mixed. With respect to mental health and substance abuse treatment, a meta-analysis of 26 studies of patient preferences for various psychotherapeutic treatments (e.g., cognitive–behavior therapy, pharmacotherapy and 12-step programs) found that, while some studies produced null findings, the data as a whole indicated a small, but statistically significant, effect on improvement across a variety of pooled mental Expert Rev. Pharmacoecon. Outcomes Res. 12(2), (2012)
Patient preferences & treatment outcomes
health outcomes (e.g., social adjustment, drug use and panic disorders) when clinicians accommodated patients’ preferences when deciding treatment [35] . One possible explanation for these findings was that patients who received their preferred treatment were also more likely to stay in treatment rather than dropping out, a claim supported by a subsequent study of patients with major depressive disorder. Kwan et al. [24] gave patients the Expectations for Treatment Inventory [36] and asked whether they preferred pharmacot herapy, talking therapy or had no preferences. Patients were then randomized to receive one treatment or the other, without taking into account their initial preferences. The authors found an indirect effect of treatment preference match on depression outcomes through its direct effect on attendance to psychotherapy sessions. In addition, whether patients stay in treatment even when clinicians satisfy patients’ treatment preferences may be moderated by the strength of the patient’s preference for that therapy. In a study of patients with major depression, Raue et al. asked patients to first rank in order their preferences for treatment (e.g., medication or psychotherapy) and then assessed strength of preference by having patients respond on five-point Likert scales to two statements about the degree to which they would like psychotherapy and the degree to which they would like medication [37] . The authors found that preference strength was a stronger predictor of initiation of treatment and adherence than was preference match per se. For biometric (e.g., blood pressure, glycated hemoglobin and cortisol level) and patient-reported physical outcomes (e.g., pain and physical function), the evidence is also mixed regarding whether accommodating patient preferences for treatment improves outcomes. A review of randomized controlled trials (RCTs) on patient preferences for treating primarily musculoskeletal conditions reported modest, but statistically significant, effects (p
