Insights from a conference on implementing comparative effectiveness research through shared decision-making

Abstract

For decades, investigators have conducted innovative research on shared decision-making (SDM), helping patients and clinicians to discuss health decisions and balance evidence with patients' preferences for possible outcomes of options. In addition, investigators have developed and used rigorous methods for conducting comparative effectiveness research (CER), comparing the benefits and risks of different interventions in real-world settings with outcomes that matter to patients and other stakeholders. However, incorporating CER findings into clinical practice presents numerous challenges. In March 2012, we organized a conference at Washington University in St Louis (MO, USA) aimed at developing a network of researchers to collaborate in developing, conducting and disseminating research about the implementation of CER through SDM. Meeting attendees discussed conceptual similarities and differences between CER and SDM, challenges in implementing CER and SDM in practice, specific challenges when engaging SDM with unique populations and examples of ways to overcome these challenges. CER and SDM are related processes that emphasize examining the best clinical evidence and how it applies to real patients in real practice settings. SDM can provide one opportunity for clinicians to discuss CER findings with patients and engage in a dialog about how to manage uncertainty about evidence in order to make decisions on an individual patient level. This meeting highlighted key challenges and suggested avenues to pursue such that CER and SDM can be implemented into routine clinical practice.

Over the past 30 years, researchers and clinicians in shared decision-making (SDM) have focused on developing and refining evidence-based methods for assessing and communicating benefits and harms of medical interventions and innovations [1–7]. The primary goal of SDM is to help patients and clinicians collaborate on health decisions by incorporating the best available evidence, as well as patients' unique characteristics and preferences [1,7,8]. A secondary goal of SDM is to reduce unwarranted practice variation in healthcare [9–11]. Decision support interventions (DESIs) based on SDM's methods and goals have been developed for a variety of health decisions [1,12,13]. DESIs aim to involve patients and family members in decisions, provide knowledge about options, help clarify preferences for risks and benefits of options, and encourage patients and family members to discuss their preferences with their clinicians.

According to the Federal Coordinating Council [101], comparative effectiveness research (CER) is “the conduct and synthesis of research comparing the benefits and harms of different interventions and strategies to prevent, diagnose, treat and monitor health conditions in real-world settings … to improve health outcomes by developing and disseminating evidence-based information to patients, clinicians and other decision-makers, responding to their expressed needs about which interventions are most effective for which patients under specific circumstances.” Traditional efficacy studies are not typically designed to inform patients and clinicians about the relative benefits and harms of one treatment compared with another for a particular patient. CER aims to answer these practical questions and, ultimately, improve the quality of healthcare delivery to real patients in real practice [14,102].

However, incorporating CER findings into clinical practice presents numerous challenges [15], including the inherent uncertainty about probabilistic information as it applies to individuals in clinical practice. Clinicians often struggle to communicate the complexity of CER data to patients in order to make informed decisions [16–18]. SDM and patient-centered DESIs could provide effective solutions for translating CER findings into clinical practice. DESIs can provide information about CER findings to help clinicians and patients choose among evidence-based options that best fit an individual patient's needs and preferences. SDM discussions between patients and clinicians can help patients make sense of research findings conducted in real-world settings with outcomes that are relevant to patients.

On 13–14 March 2012, we hosted a conference at Washington University in St Louis (MO, USA), aimed at developing a network of researchers to collaborate in developing, conducting and disseminating research about the implementation of CER through SDM. The conference consisted of two keynote speakers, four scientific abstract sessions, one poster session, six break-out discussion sessions, one-on-one mentoring opportunities and a networking lunch [103]. The keynote speakers were Michael Lauer and France Légaré. The abstract sessions included oral research presentations centered around four themes: Development and Evaluation of SDM Processes and Tools; Implementation and Impact of SDM Tools; SDM and Unique Populations; and SDM for Translating CER into Practice: the Mayo Clinic Example. The break-out discussion sessions included discussions about implementation, and training, as well as special settings and populations, including CER implementation, SDM training of clinicians, SDM with rural populations, SDM in primary care, SDM with individuals with chronic illness and SDM with elderly populations.

This article provides an overview of the highlights of the presentations and discussions that took place during the conference. The keynote addresses and abstract presentations are also available electronically [104]. We will then expand on the connection between CER and SDM and how training for CER and SDM might be integrated to focus on both developing and disseminating patient-centered research findings in routine practice.

Discussion

Conceptual similarities & differences between CER & SDM

Attendees agreed that SDM is conceptually linked to CER through translation of patient-centered research findings in routine practice. Although there are similarities in their goals to improve healthcare delivery for individuals in real-practice settings, CER and SDM differ in their underlying approach to improve healthcare delivery. The goal of CER is to generate evidence about interventions that work for specific patients under specific conditions, and to disseminate that evidence to patients and clinicians in routine practice [101]. It aims to reduce evidence uncertainty for specific patients by examining clinical effectiveness instead of focusing on evidence generated from tightly-controlled clinical efficacy studies [19]. The goal of SDM is not to generate evidence or reduce evidence uncertainty, but to communicate evidence and its associated uncertainty to patients to support patient-centered decision-making. SDM aims to provide a structured approach to decisions in which there is uncertainty about what option might work best for a given patient [8,20,21]. The deliverables produced by CER (e.g., relative effectiveness of treatments across patient centered outcomes, their relative benefits and harms for specific subgroups of patients) become key ingredients for SDM at the individual patient and clinician level. SDM, therefore, is one appropriate tool for dissemination of CER results when results of CER studies indicate that there is more than one management strategy for a clinical condition, and that choice of strategy depends on patient characteristics and preferences.

To improve clinical decision-making, many have described three levels of evidence translation: translating basic scientific data into clinical efficacy data (T1), testing the results of clinical efficacy studies in clinical practice through patient-oriented outcomes research and health services research to establish clinical effectiveness (T2), and testing the implementation of evidence to refine implementation methods and processes (T3) [15,22,23]. CER is primarily focused on obtaining evidence about patient-centered outcomes (T2 research) and testing the implementation of evidence (T3 research) according to patient-centered outcomes. SDM then employs and communicates this evidence to facilitate informed patient decisions according to patient-centered outcomes in clinical practice. As an intervention, SDM is the subject of T1, T2 and T3 research, since it is a practice based on its own basic behavioral science, including framing, risk presentation, communication and efficacy studies of its impact in controlled environments. As a process, SDM is one tool involved in good clinical practice (along with evidence-based medicine, patient-centered communication and other tools) with its ultimate goal of improving the health, functioning and quality of life of patients. Figure 1 depicts this relationship.

Figure 1. Comparative effectiveness research and shared decision-making along the evidence translation continuum.

CER: Comparative effectiveness research; EBM: Evidence-based medicine; PCC: Patient-centered communication; QOL: Quality of life; SDM: Shared decision-making.

Challenges in implementing CER & SDM in practice

Discussions during the meeting suggested that CER and SDM have faced similar implementation challenges. From a policy perspective, some have misperceived CER as a way to ration healthcare, instead of viewing it as a way to determine what interventions work best for specific subgroups of patients and improve the safety and efficiency of healthcare [24]. Similarly, some have viewed SDM as a way to convince individuals to avoid high-cost interventions, given that SDM might reduce people's interest in invasive procedures when there is uncertainty about their benefit [25,26]. To overcome these misperceptions, the group discussed incorporating key stakeholders (especially patients and members of the general public) in the design, conduct and evaluation of CER and SDM studies, and to encourage transparency about the process of generating, synthesizing and communicating evidence. Both CER and SDM leaders support these recommended strategies.

In addition, both CER and SDM face barriers because of institutional and individual practices. Meeting participants felt that many clinical guidelines or practices that are seen as standard-of-care are not always supported by the best evidence for several reasons. First, some advocacy groups and marketing practices may support policies that do not align with the best evidence [27,105]. Second, there is often fear of legal liability, leading clinicians to practice defensive medicine [28–30]. Third, individuals (including both patients and clinicians) have their own biases about best practices based on past experiences [106]. Fourth, training in and implementing CER and SDM can take up significant clinical time. Finally, both patients and clinicians can have difficulty interpreting and communicating about probabilities [31,32].

The group suggested several possible solutions to these challenges in order to facilitate CER and SDM implementation, including economic incentives, administrative support and infrastructure support. Table 1 displays some of the possible solutions discussed. For example, incorporating nurses or medical support staff prior to the physician–patient interaction could help address the time barrier to communicating about CER and engaging in SDM during clinical visits [33,34]. Mechanisms, such as training grants, could help bring a culture of CER and SDM to institutions. The overall goal of these efforts is to embed research into clinical care, train clinicians to better interpret research evidence, and train clinicians to communicate this evidence to patients through processes such as SDM. The group did not discuss ways to address issues such as legal liability per se, but engaging with key stakeholders (including policy-level stakeholders) could begin these conversations. In fact, some evidence shows that clinicians are protected when engaging in SDM with patients [29,30], and some states are implementing policies that provide additional legal protection to clinicians who engage in SDM [107].

Table 1.

Challenges and possible solutions to implementing comparative effectiveness research and shared decision-making in routine practice.

Challenges in implementation	Possible solutions
Misconceptions about the goals of CER and SDM	Engaging key stakeholders in the design, conduct and evaluation of studies Encouraging transparency about generating, synthesizing and communicating evidence
Institutional and cultural norms do not support CER and SDM	Administrative and infrastructure support for CER and SDM (e.g., incorporating nurses or medical support staff prior to the physician–patient interaction)
Clinicians need time and resources to learn about CER and SDM skills	Economic incentives for clinicians to learn and implement CER and SDM Institutional training grants and resources available to teach clinicians about interpreting research evidence, analyzing administrative data, analyzing effectiveness data, communicating probabilities and providing decision support to patients Embed SDM and CER training into existing programs (e.g., as a supplemental module to a training program in evidence-based medicine) Videos to view SDM processes and skills in action

CER: Comparative effectiveness research; SDM: Shared decision-making.

Many participants discussed the need both to work within existing educational frameworks and build incentives for clinicians to learn about and engage in SDM and CER translation. However, because policy level changes can take years to implement, participants suggested starting by embedding CER and SDM training into existing frameworks that clinicians already know (e.g., patient-centered care and evidence-based medicine). The content of SDM training as a way to communicate CER evidence should include general communication skills training [33,35], as well as training in understanding, interpreting and communicating probabilities according to latest standards in risk communication [2,32]. Training materials should use language that clinicians recognize and yet help them distinguish SDM from related concepts. SDM is part of an approach that places patients and their concerns at the center of its action and purpose. This approach, often called patient-centered care, extends beyond the way in which decisions are made to include all aspects of the care [36]. When clinicians seek to move patients into adopting or changing a behavior (e.g., begin exercising or quit smoking), other forms of patient-centered care are more suitable, such as motivational interviewing [21]. Unlike motivational interviewing, however, SDM shares information with patients based on the best available evidence in situations with no dominant clinical option. Evidence alone cannot fully inform a clinical decision in these situations, and patient values, goals, preferences and context are important. A clinician that is able to implement SDM in these situations, therefore, is providing patient-centered care and practicing evidence-based medicine. In addition to discussing related concepts, trainings could use videos demonstrating SDM compared with standard of care to help clinicians hear what SDM actually sounds like in practice [108]. Dr Légaré encouraged those providing training in SDM to focus on multifaceted interventions, to emphasize modifiable SDM behaviors and skills, and to consider activating healthcare professionals, patients and policy members to support SDM [37,38].

Incorporating SDM with specific populations

Although SDM works at the individual level, many DESIs and models of patient engagement have focused on `typical' patients. Given that the focus of CER is in defining effectiveness for patient subgroups and individuals, we engaged in discussions regarding some populations for whom SDM may need to be further tailored and refined.

SDM with rural populations

There was a rich discussion about SDM with rural populations focusing on characteristics that make rural populations unique. Some of these challenges also provide key opportunities to impact rural healthcare through engaging in SDM. Rural populations in different states and regions in the USA can have different cultural norms, values, preferences and healthcare needs. However, one common challenge is the difficulty of traveling to/from tertiary care centers when a health condition warrants specialized care. Thus, leveraging technology to provide information about decisions in advance of appointments could reduce travel time to/from clinics for multiple appointments and can help start the process of SDM [39]. There are numerous patient-centered DESIs that can be available for viewing online that can help patients living in rural areas to learn about their illness, clarify their values about options available [25], and could even help patients engage in online support through connecting to others who have been through similar decisions in the past. These online opportunities can help prepare patients for SDM discussions with their clinicians and can reduce the likelihood that they will need multiple appointments before making a decision. In addition, some suggested that these online resources could help normalize the illness experience and the process of engaging in SDM for those living in rural communities.

SDM with elderly populations

Given that elderly patients might be more familiar with paternalistic styles of communication in healthcare [40], attendees recommended that clinical discussions begin by assessing patients' desired level of involvement in the decisions confronting them (as decision-making preferences can vary across health conditions and circumstances [41]). Furthermore, CER studies are essential in order to clarify how evidence might apply to elderly patients, as the elderly are often excluded from traditional randomized clinical trials [42,43]. SDM should focus on both quality of life and survival benefits for elderly patients. In addition, older adults are a heterogeneous group with respect to both cognitive and physical functioning [44], thus there might not be universal guidelines for clinical decision-making by age alone. Attendees suggested borrowing language from palliative care settings when discussing treatment decisions with the elderly so that discussions are not perceived as `rationing' care. For example, for elderly patients who are considering forgoing aggressive chemotherapy when the risks could outweigh the benefits, rather than suggesting `doing nothing' as an option, clinicians could use the phrase `choosing to prioritize symptom control'. Further research on engaging in SDM with surrogate decision-makers for those who have extensive cognitive impairment, could also benefit elderly populations.

SDM with individuals with communication disorders

Individuals with communication disorders (such as motor speech disabilities that do not necessarily affect cognitive or language abilities) often report dissatisfaction with communication with their providers and face obstacles to receiving high-quality care as a result [45–47]. A qualitative study about the health decision-making experiences of individuals with communication disorders revealed several important suggestions for engaging in SDM. A qualitative study of individuals with communication disorders' experiences with health decisions, revealed several important suggestions for engaging in SDM with this population [109]. Before engaging in SDM, clinicians can initiate discussions by asking how best to communicate with these patients. Next, clinicians can begin the SDM process by establishing patients' preferred level of involvement in decisions and patients' preferred approach to involving caregivers (who are often present at health appointments) in health discussions. This process can help prevent clinicians' misperceptions that a caregiver is the primary decision-maker or that patients with communication disorders are cognitively incompetent to make decisions. Additional research should be conducted to evaluate the roles of patients, clinicians and caregivers in these unique decision-making contexts.

Examples of successful implementation of CER & SDM in practice

Where present, SDM implementation often exists as experimental or pilot interventions, or as activities carefully deployed on the boundaries of care [13]. Since 2005, Mayo Clinic (MN, USA) investigators have been developing DESIs [48–52] and implementing them for evaluation in usual clinical settings that include primary care practices in rural, suburban and urban settings, specialty clinics, hospitals and the emergency department. Several abstract sessions discussed the Mayo Clinic's implementation efforts.

The Mayo Clinic's DESIs support empathic decision-making about chronic conditions [53], such as the choice to start on preventive medications (e.g., statins, bisphosphonates and the myocardial infarction bundle), decisions about multiple available drugs to treat chronic conditions (e.g., antidiabetic and antidepressant agents), and decisions about available courses of action (e.g., admission and urgent testing or discharge and delayed evaluation of low-risk chest pain in the emergency department). The DESIs are designed for use during clinical consultations, which requires attention to the dynamic and contextual features that can hinder or promote SDM [13]. Patients and clinician stakeholders participated in the DESI development team and in the cocreation of tools that could `create conversations' about decision-making. In each of the cases, the tools were implemented outside of the electronic health record, in part because of design considerations and in part because of the difficulties inherent with competing against other priorities of the information technology departments. It became clear, however, that integration into the workflow of these practices was essential: an implementation in the Mayo Clinic health record of the Statin Choice decision aid gathered more than 2000 uses without advertising, physician training or incentive. To date, Statin Choice and Diabetes Medication Choice have both become embedded into the electronic health record at the Mayo Clinic.

Overall, the efficacy-tested Mayo Clinic DESIs have shown the ability of well-designed tools to translate evidence from CER reviews into patient-centered tools [54–60]. The trials of these tools show the feasibility of DESIs deployed at the point-of-care when designed for this purpose. The outcomes obtained thus far suggest that many of these tools (but importantly not all) are efficacious in primary and secondary care, and in hospital and emergency care. The mean age across these trials has varied from 62 to 67 years, suggesting that older age is not a barrier for DESIs or SDM. Clinician satisfaction has ranged from 74–90%, and the mean incremental time investment in using these tools has ranged from 2.5 to 3.8 min (with a wide range in each study). Analysis of video recordings shows clinicians with minimal training delivering the majority of the elements of the DESI properly. Across these trials, DESI implementation is associated with a 20% improvement in patient knowledge about options, and a 17% increase in clinicians' effort to engage patients in decision-making. The impact on health outcomes, decisions and adherence remains heterogeneous and inconclusive [54–60].

Detailed implementation work on >50 clinical sites, >200 clinicians and >1000 patients thus far offers evidence that SDM and DESIs are feasible and useful innovations that complement the investments in evidence generation and help realize its value at the point-of-care [61]. Although the conference focused on these examples as ways to translate CER through SDM, there are many DESIs that were not explicitly discussed at our conference, and results of those can be accessed in the latest Cochrane review of DESIs and outcomes [25].

Conclusion & future perspective

Meeting attendees agreed that in order to implement SDM and translate CER into practice, engaging key clinician, patient and policy level stakeholders is of utmost importance. This process is consistent with the Patient Centered Outcomes Research Institute's methods that incorporate key stakeholders while examining and promoting the best evidence. In the short term, relying on existing infrastructure can help move SDM into practice. Eventually, policy-level changes and incentives that can work within the current system are needed to make SDM more likely to be sustained; however, these incentives can take years to implement. The Mayo Clinic's implementation efforts provided an example of ways to work within the current system in order to translate evidence from CER reviews into patient-centered tools.

Attendees also agreed that using language familiar to clinicians and patients can help generate more support for SDM and more willingness to learn about the SDM process. However, many cautioned that it is important to maintain conceptual clarity and differentiate SDM from other related processes, such as motivational interviewing, evidence-based medicine and patient-centered communication. Attendees agreed that SDM training should focus on the multiple components of the SDM process, including content, such as communication skills training, training in interpreting and discussing probabilities, and training in the use of DESIs.

Meeting attendees suggested several areas for future research, including the role of surrogate decision-makers in SDM [109] and the process of engaging in SDM with elderly populations who might have cognitive limitations. They also suggested continuing to incorporate patient safety outcomes in studies, such that there is evidence to demonstrate that patients are not worse off from a health or quality of life perspective because of engaging in SDM. In addition, they encouraged researchers to consider whether SDM protects clinicians from malpractice suits given that one barrier to engaging in SDM is fear of legal liability [29,30]. Finally, they suggested research on SDM as used across teams of clinicians since many health decisions are complex and are often made across a series of conversations [33,62,63].

CER and SDM are related processes that emphasize examining the best clinical evidence and how it applies to real patients in real practice settings. SDM can provide one opportunity for clinicians to discuss CER findings with patients and engage in a dialog about how to manage uncertainty about evidence to make informed decisions on an individual patient level. This meeting highlighted key challenges and suggested avenues to pursue such that CER and SDM can be implemented into routine clinical practice.

Executive summary.

Background

• ■Shared decision-making (SDM) and patient-centered decision support interventions could provide effective solutions for translating comparative effectiveness research (CER) findings into clinical practice.
• ■On 13–14 March 2012, we hosted a conference at Washington University in St Louis (MO, USA) aimed at developing a network of researchers to collaborate in developing, conducting and disseminating research about the implementation of CER through SDM.
• ■This article provides an overview of the highlights of the presentations and discussions that took place during the conference. We then expand on the connection between CER and SDM and how training for CER and SDM might be integrated to focus on both developing and disseminating patient-centered research findings in routine practice.

Discussion

• ■Attendees agreed that SDM is conceptually linked to CER through translation of patient-centered research findings in routine practice.
• ■The deliverables produced by CER (e.g., relative effectiveness of treatments across patient-centered outcomes, their relative benefits and harms for specific subgroups of patients) become key ingredients for SDM at the individual patient and clinician level.
• ■CER and SDM have faced similar implementation challenges.
• ■Several possible solutions to these challenges were suggested in order to facilitate CER and SDM implementation, including economic incentives, administrative support and infrastructure support.
• ■In addition, given that the focus of CER is in defining effectiveness for patient subgroups and individuals, we engaged in discussions regarding some populations for whom SDM may need to be further tailored and refined.
• ■Several abstract sessions discussed the Mayo Clinic's (MN, USA) implementation efforts as one example of successful implementation of CER through SDM.

Conclusion

• ■Meeting attendees agreed that in order to implement SDM and translate CER into practice, engaging key clinician, patient and policy-level stakeholders is of utmost importance.
• ■Using language familiar to clinicians and patients can help generate more support for SDM and more willingness to learn about the SDM process.
• ■Several areas for future research were discussed.