By Stacey L. Sheridan, M.D., M.P.H.,a,b Russell P. Harris, M.D., M.P.H.,a,b Steven H. Woolf, M.D., M.P.H.c for the Shared Decisionmaking Workgroup of the United States Preventive Services Task Force.
Address correspondence to: Stacey L. Sheridan, M.D., M.P.H.; The Cecil G. Sheps Center for Health Services Research; 5039 Old Clinic Building, CB 7110; University of North Carolina at Chapel Hill; Chapel Hill, NC 27599. E-mail: ssherida@email.unc.edu
This article originally appeared in the American Journal of Preventive Medicine. Select for copyright, source, and reprint information.
Introduction
Patient-Clinician Interactions on Screening and Chemoprevention
What Is Shared Decisionmaking?
Why Engage Patients in Shared Decisionmaking?
Do Patients Want to Be Actively Engaged in Shared Decisionmaking?
Are Patients Able to Engage in Shared Decisionmaking?
What Are Barriers to Clinician Use of Shared Decisionmaking?
How Might Clinicians Facilitate Patient Participation in Decisionmaking?
Conclusions
Acknowledgments
References
Author Affiliations
In 1984, the Department of Health and Human Services established the U.S. Preventive Services Task Force (USPSTF) as an independent panel of nonfederal experts that would develop evidence-based recommendations on clinical preventive services based on systematic reviews of published research and explicit decision rules for translating science into practice policy. At that time, the central question was whether there was high-quality evidence showing that a preventive service improved health outcomes.
Recommendations were graded according to a scheme adapted from the Canadian Task Force on the Periodic Health Examination, which focused on whether evidence existed to support performing the service as part of the periodic health examination (Figure 1A). Most preventive services had not been formally studied, and the USPSTF gave many "C" recommendations, indicating insufficient evidence was available to recommend for or against the service. Interventions such as mammography and hypertension screening, however, received "A" recommendations because formal studies showed that they improved health outcomes. Implicitly, clinicians were encouraged to promote preventive services that had been shown to improve outcomes and give lower priority to those with unevaluated health effects.
In the years that followed, multiple studies of prevention appeared in the literature, enabling the USPSTF to recommend a larger number of preventive services. Because these studies included more precise data on the magnitude of potential benefits and harms than had previously been available, the USPSTF was faced with answering a new and more complex question: is the magnitude of potential benefit from a service sufficient to outweigh the magnitude of potential harm?
In time, the USPSTF encountered preventive services for which even this question could not be easily answered. The answers did not jump from the pages of clinical trials but instead depended on the value that individual patients assigned to the potential harms and benefits. Although in some cases the USPSTF was comfortable weighing those tradeoffs based on assumptions of how most patients would value the outcomes, increasingly the USPSTF recognized that patient preferences were too variable to reach a generic determination of what was best.
To reflect its growing recognition of the complexity of its decisionmaking, the USPSTF published a new framework for grading recommendations (Figure 1B).1 This framework drew a distinction between the USPSTF's assessment of the quality of evidence and its subjective judgment about the degree to which benefits outweighed harms. Under this new scheme, the USPSTF assigned "C" recommendations to services thought to have small net benefit when averaged across the population. The USPSTF recognized, however, that the net benefit might be larger for population subgroups with special risk factors or for individual patients with personal preferences that differed from those of the panel. Given that recognition, the USPSTF decided not to make a generic recommendation for "C" services, but rather to highlight that all "C" decisions are "likely to be sensitive to individual patient preferences."1
Strength of Recommendations A: There is good evidence to support the recommendation that the condition be specifically considered in a periodic health examination. B: There is fair evidence to support the recommendation that the condition be specifically considered in a periodic health examination. C: There is poor evidence regarding the inclusion of the condition in a periodic health examination, but recommendations may be made on other grounds. D: There is fair evidence to support the recommendation that the condition be excluded from consideration in a periodic health examination. E: There is good evidence to support the recommendation that the condition be excluded from consideration in a periodic health examination. |
| Net Benefit | ||||
| Quality of Evidence |
Substantial | Moderate | Small | Zero/Neg |
|---|---|---|---|---|
| Good | A | B | C | D |
| Fair | B | B | C | D |
| Note: Poor = I, insufficient evidence to determine the net benefit of the service or to recommend for or against routinely providing it. | ||||
Explicit language encouraging consideration and discussion of patient preferences also began creeping into the rationale for those "A" and "B" recommendations in which the USPSTF recognized large or moderate net benefit, but also a substantial potential for harm or a significant amount of uncertainty about which screening modality was best. For instance, in 2002, the USPSTF issued a "B" recommendation for mammography screening and also recommended a conversation with women to ensure a realistic understanding of the limited absolute benefit of screening in the face of potential harms.2 Additionally, recommendations for colorectal cancer screening indicated that patients should be invited to participate in the process of determining which of the 4 available screening options was preferable.3
This trend toward recommending discussions to elicit patient preferences introduced some discomfort for the USPSTF. The USPSTF did not want to suggest that informing patients about the benefits and harms was unnecessary for services in which discussions were not explicitly recommended; rather it envisioned a more systematic process for some services. Members of the Task Force disagreed about whether such discussions should be practiced for all preventive services or should be advocated for only certain services, such as those involving "close calls." Many Task Force members expressed concerns for the busy clinician who lacks the time to engage in extended discussions about every potential clinical preventive service that a patient might want. As confusion began to surround the question of whether "A" and "B" recommendations from the USPSTF meant that the preventive service was to be "performed" or "discussed," the panel decided to formulate some suggestions about how the clinician might approach discussions about screening and chemoprevention.
These USPSTF suggestions are summarized in this article, along with commentary on the current thinking and evidence regarding shared decisionmaking between patients and clinicians. Unlike conventional USPSTF reports, this document is neither a systematic evidence review nor a formal recommendation statement. The USPSTF comes to this topic not with its customary objective of evaluating effectiveness, but rather to articulate its finding that shared decisionmaking is a necessary tool for making recommendations to individual patients concerning interventions that have net benefit for some but not for others. This article is, therefore, a concept paper, intended to clarify how the USPSTF envisions the application of shared decisionmaking in the execution of preventive services recommendations.
The suggestions herein were derived from a non-systematic evidence review and an iterative dialogue between the authors of this paper and USPSTF members, experts, clinicians, and representatives from the Task Force on Community Preventive Services at the Centers for Disease Control and Prevention. These suggestions underscore the USPSTF's growing recognition that some recommendations need to be individualized according to patients' special circumstances and preferences.
The USPSTF does not endorse a specific style of patient interaction, but does have suggestions for clinicians on how to interact with patients for each of its 5 categories (A, B, C, D, I) of screening and chemoprevention recommendations.
The USPSTF encourages clinicians to inform patients about recommended services. Ideally, this means that clinicians track the "A" and "B" recommended preventive services for each patient, making sure that patients are informed that these services are recommended at given intervals. Many patients may already be informed about these services or accept them as a routine part of the medical exam (e.g., screening for high blood pressure or obesity), making lengthy and frequent discussion unnecessary. It is prudent, however, for clinicians to be prepared to discuss the potential benefits and harms of these services if patients indicate an interest. For example, clinicians might sensibly prepare to respond to patients' concerns about the rates and consequences of false-negative Papanicolaou smears during cervical cancer screening.
For some "A" and "B" recommended services, clinicians may want to consider more discussion. For example, the USPSTF has recommended (A recommendation) that clinicians discuss aspirin chemoprevention with adults at increased risk for coronary heart disease and that adults not take aspirin prophylaxis without understanding the likelihood of specific benefits and harms.4 Patients with different values might make different decisions about taking prophylactic aspirin, thus discussion of preferences is necessary. Similarly, the USPSTF has recommended (B recommendation) engaging women who are at high risk for invasive breast cancer and low risk for the adverse effects of chemoprevention in discussions to determine whether chemoprevention for breast cancer is appropriate.2 This decision also cannot be made from existing evidence but depends on personal preferences.
A proactive approach to discussions may also be desirable for certain "C" and "I" recommended services, particularly if they have high visibility due to substantial media attention and may be on the patient's mind (e.g., prostate cancer screening) or if they have the potential to substantially affect a given patient (e.g., ovarian cancer screening in a woman who has multiple first-degree relatives with ovarian cancer). Given the demands of clinical practice, however, the USPSTF encourages clinicians to rely on clinical judgment when deciding which discussions of "C" and "I" recommendations to initiate and to consider decision aids or trained assistants to help provide information about these services. Community approaches such as those suggested by the Task Force on Community Preventive Services5 may also be helpful to patients in understanding the trade-offs involved with "close call" or "uncertain" services.
Clinicians are generally under no obligation to initiate discussions of services with "D" recommendations; these are services that the USPSTF has found to be either of no benefit or potentially harmful. Nonetheless, clinicians should be prepared, with the help of decision aids and trained assistants, to explain why these services are discouraged. Clinicians should also consider a proactive discussion for services with high visibility or special importance for the individual, or for services for which new evidence has prompted withdrawal of previous recommendations. For instance, with recent evidence demonstrating the adverse effects of hormone replacement therapy (HRT) on the heart,6 clinicians should consider raising the issue of the appropriateness of HRT with the many women who are currently receiving such chemoprevention. Clinicians may also want to proactively raise the issue of HRT with a woman who places special value on reducing the risk for bone fractures and wants to consider all options. In such circumstances, the woman should be fully informed of the tradeoffs between potential benefits and harms and of her unique situation that could dictate departure from recommended practice.
Whenever decisions about preventive services must be made, the USPSTF encourages informed and joint decisions. This means that patients should be informed about preventive services before they are performed and that the patient-clinician partnership is central to decisionmaking. The need for this is most clearly demonstrated for those "C" and "I" recommendations that require decisionmaking because of high visibility or special importance to the individual. For such "C" recommendations, in which the average net benefit is small (e.g., osteoporosis screening in postmenopausal women under the age of 60 or aged 60-64 with no risk factors), patients may be aided not only by evidence-based information about the magnitude of benefit and its close tradeoff with potential harms, but also by clinician assistance in determining whether their individual risk profile and personal preferences make the net benefit positive or negative. Similarly, for "I" recommendations, in which the evidence is insufficient to recommend for or against screening (e.g., prostate cancer screening), patients may sometimes need both discussion of the uncertainty that precludes a clear recommendation and clinician assistance in determining their preference for or against action in the face of uncertainty.
One important form of informed and joint decisionmaking is "shared decisionmaking," in which patients are involved to the extent that they desire as an active partner with the clinician in clarifying acceptable medical options and in choosing a preferred course of clinical care. Although available research shows mixed results about the effect of this type of decisionmaking on health outcomes, support for a patient-clinician partnership in decisionmaking comes from a combination of ethical and practical arguments, which are detailed below.
Decisionmaking within a patient-clinician partnership has been alternately called "shared decisionmaking,"7-9 "informed decisionmaking,"10-11 "informed shared decisionmaking,"12 "evidence-informed patient choice,"13-14 "patient-centeredness,"15 "enhanced autonomy,"16 "relationship-centered decisionmaking,"16 "deliberative decisionmaking,"17 "interpretive decisionmaking,"17 and "mutual participation."18 Whatever its name, decisionmaking within the patient-clinician partnership universally encompasses a process in which both the patient and clinician share information with each other, take steps to participate in the decisionmaking process, and agree on a course of action.
In some models, the process has clearly defined steps or competencies that have been identified through focus groups or literature reviews.9-16 In other models, the process is left to the discretion of the individual patient and clinician.7,8,17,18 Of those models that define specific steps, many acknowledge the patient's right to relinquish the decision to the clinician and proceed in a paternalistic model,7-10,12,13 several call for evidence-based presentations of information,10-14 a few call for physicians to express their preferences,7,16,19,20 and a few call for an explicit check of patient understanding.9,10 One model acknowledges the practical limitations of medical practice,10 proposing a hierarchy of decision complexity, with more complex decisions requiring a greater intensity of interaction than simple decisions.
Some confusion has surrounded the use of various terms for decisionmaking within the patient-clinician partnership, particularly the terms shared decisionmaking and informed decisionmaking. Researchers,7 patients, and clinicians sometimes use the term informed decisionmaking to describe both independent decisionmaking by the patient and joint decisionmaking by the patient and clinician.
The USPSTF, in collaboration with the Task Force on Community Preventive Services,5 defines informed decisionmaking (IDM) as an individual's overall process of gathering relevant health information from both his or her clinician and from other clinical and non-clinical sources, with or without independent clarification of values. The Task Force defines shared decisionmaking as a particular process of decisionmaking by the patient and clinician in which the patient:
This process has the goal of an informed and joint decision. Thus, although the definition focuses primarily on evidence for patient involvement, the process necessarily requires clinicians to reveal their clinical reasoning and biases to facilitate a truly joint decision.
Shared decisionmaking differs significantly from decisionmaking under the doctrine of informed consent, which arose in the law in the mid-1970s. Although informed consent was an obvious forerunner of shared decisionmaking, its focus in practice was on clinician disclosure rather than on joint participation. Informed consent did mandate that patients actively express consent rather than just expressing agreement with, yielding to, or complying with proposed medical care.21 Its success was measured, however, by clinician disclosures before risky procedures. These disclosures included a description of the proposed treatment, the alternatives to the proposed treatment, and the inherent risks of death and bodily injury, as well as any other information that a reasonable clinician would disclose or a reasonable patient would want to know in the same circumstance. Clinicians could forgo these disclosures if a patient did not want to be informed, if the procedure was simple with little risk, and, if (in the clinician's judgment) it was not in the patient's best interest to know.22 Although patients' signatures signified their receipt of the information, little attention was given to ensuring active involvement in decisionmaking.23
Shared decisionmaking also differs from decisionmaking in consumerism, in which patients gather relevant information, which they obtain from their clinician and other sources (e.g., the Internet), and independently determine which options they prefer. In this model, the clinician's role is not to serve as a partner in decisionmaking; rather it is to execute the selected intervention, without giving special attention to exploring the patient's understanding, values, or reasoning.
Over the last few decades, interest in shared decisionmaking has been growing. In a recent report, a committee of the Institute of Medicine suggested that "a patient-provider partnership is needed to ensure that decisions respect patients' wants, needs, and preferences and that patients have the education and support they require to make decisions and participate in their own care."24 To highlight the importance of this partnership, the committee identified creating a patient-provider partnership as 1 of the 6 principal aims of the 21st-century health care system.
Shared decisionmaking can be recommended on multiple grounds. From an ethical perspective, it promotes patient autonomy, protecting the integrity of the patient as an independent and rational decisionmaker capable of self-determination. From the interpersonal perspective, shared decisionmaking promotes trust in the patient-clinician relationship25 and may enhance the confidence of patients to participate in their health care.26 From an educational perspective, shared decisionmaking improves knowledge about screening and chemoprevention options, creates more realistic expectations about benefits and harms, and reduces the decisional conflict associated with feeling uninformed.27,28 From a utility perspective, the "best choice" for decisions involving close tradeoffs can only be made by incorporating the personal preferences of the patient.
From a health perspective, evidence that shared decisionmaking improves health outcomes is indirect and mixed. Systematic reviews of decision aids, which are based on the tenets of shared decisionmaking and provide patient education and values clarification, have shown no consistent demonstrable effect on health outcomes.27-32 This result might have been anticipated because decision aids help patients choose among alternatives in which the balance of benefits and harms is a "close call," thereby creating an environment in which the net health benefit across a population of reasonable people who choose differently might approach zero.
Only a few studies33,34 have measured adherence to a chosen course of action; such measurements could give further insight into the effects of decision aids and shared decisionmaking. Interventions that have provided patients with training in information-seeking and negotiation skills have resulted in improvements in symptoms and physiologic outcomes.26,35-38 Importantly, these studies have targeted health conditions in which benefits generally outweigh harms for most individuals. Studies in which patients perceive that they negotiate a common plan with the clinician also show benefits in health outcomes,19 as do studies in which patients perceive that they are active participants in decisionmaking (regardless of their preferred role).39-41
Patient willingness to participate is critical for shared decisionmaking. A recent review42 reported mixed interest in participation (19-68 percent), noting that patients who were younger and more highly educated showed greater interest. The authors offered possible explanations for mixed interest, including lack of a clear distinction between medical problem solving, which requires tasks for which patients are not qualified, and medical decisionmaking, which could be shared.43-44 The authors also noted that patients may not realize that medicine is an inexact science, believing that their clinicians prescribed the only treatment available. Other reasons for lack of interest in participation include lack of understanding that there is a decision to be made, discomfort with a new role in decisionmaking, inexperience with clinicians employing this approach to decisions, steadfastness in preconceptions about the course of care,43 membership in an ethnic group that does not value patient autonomy,45,46 and fear of regret for decisions that turn out badly. Clinicians should address such concerns and misconceptions about participation in decisionmaking before assessing patients' desire to participate.
Even when patients are interested in shared decisionmaking, inability to understand medical concepts may limit their participation. For instance, multiple studies have demonstrated that some patients have difficulty understanding risk concepts,47-54 which calls into question their ability to accurately weigh the benefits and harms of preventive services. Low functional literacy and numeracy, each of which affects approximately a quarter of the U.S. population,55 exacerbate these difficulties. Low literacy has been repeatedly associated with reduced health knowledge and poor outcomes.56,57
Even patients who have good reading and numeric skills may find themselves at a loss in medical conversations. Clinicians frequently speak in a medical jargon that is inaccessible to patients. Clinicians also use ambiguous qualitative descriptions such as "some" or "likely," which patients may interpret differently than clinicians intended.58,59 Furthermore, the normal constraints of short-term memory limit patients' ability to walk away from in-depth discussions remembering all key information.60
Fortunately, the number of resources available to address these problems is increasing.56,60,61 These resources allow patients to privately consider complex medical information over time.29 That said, it is unclear whether patients need to comprehend and remember large volumes of complex information to share decisions about screening and chemoprevention with their clinicians. The effectiveness of providing simple, focused information in concise formats to facilitate shared decisionmaking requires further study.
Clinician interest is also of critical importance to shared decisionmaking, but few studies have expressly examined clinician interest in engaging patients in the decisionmaking process.42 Many have observed, however, that clinicians currently face barriers that could potentially diminish initiation of shared decisionmaking.42,62
Even the most well-intentioned and conscientious clinicians have difficulty in engaging patients in all appropriate preventive care. With shortened office visits, primary care clinicians struggle against competing demands and opportunities to find the time for prevention.63-66 Including patients in decisionmaking may aggravate these struggles, resulting in reduced or delayed action as patients consider their options. Alternately, shared decisionmaking may provide patients with skills that improve decisionmaking and motivation across other aspects of their health care,67 freeing up more time for prevention. The long-term effects of shared decisionmaking on competing clinical demands are hard to estimate and ripe for study.
The potential financial costs of shared decisionmaking are also of concern. Lack of financial reimbursements is a disincentive to take the time to discuss topics such as breast or colon cancer screening. Additionally, the cost-effectiveness of such discussions has not been studied. We expect the cost-effectiveness of shared decisionmaking to be best for decisions that are highly sensitive to patient values.68 These decisions could be identified by formal decision analysis,69 but more likely correspond to decisions for which the USPSTF has felt uncomfortable making global recommendations. A hierarchical approach to shared decisionmaking (ie, using shared decisionmaking for some, but not all, decisions), as suggested in this article, might be expected to maximize the cost-effectiveness of shared decisionmaking for preventive services.
Even when clinicians have the necessary time and financial support for shared decisionmaking, additional barriers exist. Many clinicians lack training in the interviewing techniques needed to engage patients in decisionmaking.70 The lack of accurate, organized scientific evidence about the benefits and harms of many preventive services leads to confusion about the potential benefits and harms of screening and chemoprevention services, making shared decisionmaking more complex. When organized evidence is available, providers are often uncertain about which decisions require patient participation and about how to communicate technical concepts to patients in simple language that is accurate, balanced, and understandable.
With these barriers, clinicians, regardless of their interest, may have difficulty engaging patients in shared decisionmaking. Although the potential of creative solutions to increase shared decisionmaking is uncertain, system-level changes may be worth exploring. Health plans could alter reimbursement to encourage time for shared decisionmaking. Delivery systems could modify visit schemes to facilitate group education or promote the use of nurses or case management teams to relieve the time burden for clinicians. Health plans and delivery systems could post shared decisionmaking materials on Web sites providing decision support to both patients and clinicians. Such interventions have been shown to improve the processes and outcomes of care in chronic illness71,72 and offer promise for surmounting barriers to informed and joint decisions.