Implementing shared decision-making: consider all the consequences
Elwyn, Frosch, Kobrin · 2016
공개 학술지 Implementation Science : IS 에 실린 총설 논문입니다(2016년). CC BY 4.0 으로 공개된 글이고, 그림 · 표 · 참고 문헌 표시를 뺀 본문을 실었습니다.
출처 Elwyn et al. (2016), Implementation Science : IS · 라이선스 CC BY 4.0 · 원문
이 장은 원문만 있습니다. 문장별 해석은 앞 장부터 차례로 붙입니다.
2장 · Implementing shared decision-making: consider all the consequences — Approach
To achieve our goals, we took the following approaches:
• We considered published work that has examined outcomes relating to shared decision-making. We focused on systematic reviews and other high-quality narrative reviews.
• We identified research gaps highlighted by existing reviews and discussed the preliminary results of our work at academic meetings and conferences.
• We developed a multilevel framework to describe potential consequences over time and propose a sample of questions to illustrate potential future research directions.
To do the work, we examined a review of outcomes considered in existing studies of shared decision-making and analyzed the contribution of patient decision aids. In addition, we reviewed other relevant systematic reviews, including an assessment of whether patient decision support had led to cost savings, the effect of shared decision-making on health inequalities, on litigation, and on treatment satisfaction and clinical outcomes. After reading these reviews, we met, face-to-face and virtually, on multiple occasions. Between meetings, we used a cloud-based document editing system to develop a figure, tables, and manuscript. A draft version of the framework was submitted to a relevant conference, where feedback shaped a penultimate version. We imagined a range of possible consequences, considering they might be good for some but not for others, recognizing that different consequences might not be consistently valued by different stakeholders. As we developed the framework, we decided to group outcomes into those that were proximal, distal, and distant to a shared decision-making process.
We categorized outcomes by thinking about the level of impact. Figure 1 provides an overview of a process where collaborative deliberation leads to an series of hypothesized effects. Table 1 provides more detail. We did not wish to impose definitive time limits because we acknowledge that divisions will not be precise, and some consequences might accrue gradually. Parsing potential outcomes into these time frames enabled us to consider the consequences of collaborative deliberation that go beyond those that are obvious or easily measured.
Main argument
Many measurement approaches to shared decision-making have been developed. Shay and Lafata noted that “the empirical evidence” about the effectiveness of shared decision-making has not been “systematically summarized” and set about the task. Shay and Lafata found three sets of reported measures: (1) cognitive-affective, (2) behavioral, and (3) health outcomes, reporting that patient “health outcomes were the least studied.” There were no efforts made to suggest other types of consequences.
Significant efforts to implement patient decision aids have taken place, mainly in the USA and in the UK. The hospital system at Dartmouth Hitchcock was the first center dedicated to delivering patient decision aids, and more than a decade later, it continues to offer such tools to some categories of patients. More recently, Group Health in Seattle implemented a number of patient decision tools alongside a quality improvement initiative aimed at surgical specialists. Evaluations at Group Health demonstrated high levels of patient exposure to the tools, higher than those achieved in other organizations, and evidence of impact on surgical rates. Others have reported difficulties in attempting to introduce these tools into organizations. Reviews of efforts made to implement patient decision aids indicate how difficult it is to alter routine workflows. Most of the research however has focused on how to provide decision aids to specific patient groups before they attend clinic or on how to motivate clinicians to provide decision aids to relevant patient groups. Much less attention has been given to research that would identify organizational, system, and policy factors that could influence practice change.
Few of these implementation-type studies have considered the effects of shared decision-making, beyond the proximal outcomes of decision support interventions given to patients. The work at Group Health is a notable exception; their positive results serve to illustrate what might be possible. Even so, they did not evaluate a fuller range of consequences or over a time horizon. It is not therefore possible to conclude much at all about the mid or longer-term consequences of a concerted effort to undertake shared decision-making at a system wide level, whether defined as decision aid use or as a communicative process of collaboration and deliberation. A handful studies have included costs and cost-effectiveness, but these have all been related to the evaluation of patient decision aids. The results of a systematic review of studies with economic assessments revealed the need for more and higher quality studies before firmer conclusions could be drawn. A systematic review of shared decision-making in relation to litigation found five studies, mostly case reports, and was unable to reach a firm conclusion.
Conceptual framework
In this section, we provide a figure which illustrates a hypothesized set of proximal, distal, and distant consequences of collaboration and deliberation undertaken routinely (see Fig. 1), proposing that well-informed preference-based patient decisions may lead to safer, cost-effective healthcare, which in turn reduce utilization rates and improve health outcomes.
To illustrate the potential positive, negative, and potentially peverse consequences, we provide more detail (see Table 1), where we describe a range of potential consequences, with commentary below.
Individual consequences—patient and clinician level
As Shay et al. noted, outcomes at the individual level have been those most commonly considered, where the majority of interventions have been decision aids given to patients before clinical encounters and the effects evaluated soon afterwards. It is reassuring to see that such tools show that patients develop preferences based on their comprehension of accessible information, have more realistic expectations, less “decisional conflict” as currently measured, and greater satisfaction. Emerging evidence also suggests that shared decision-making can improve adherence and increase trust. It is also reassuring that these tools do not cause significant anxiety or other problems for patients.
That these tools often reduce the intention to choose intensive treatments has led to most speculation about the prospect of modified resource use and possible cost reduction at distal and distant time points. Nevertheless, this outcome is understudied. Because of the need for long-term follow-up, studies that use modeling methods and sensitivity analyses might be helpful. We know of no qualitative studies that investigated the reasons for the lowered interest in some treatments.
Individual level effects on clinicians should also be considered; the experience of supporting patients in arriving at informed decisions may be intrinsically rewarding. Clinicians may also find the effort involved emotionally and cognitively taxing, adding to their workload burden. These consequences for clinicians should be evaluated and understood as they should be expected to influence the uptake of shared decision-making.
Interactional and group level consequences
Few studies have examined the impact of shared decision-making as a communicative process beyond the effect on patients. Experts in organizational psychology consider five types of potential outcomes: (1) tangible outcomes products of teaming, e.g., costs or rates; (2) attitudes or emergent states, e.g., trust, psychological safety; (3) cognitive states, e.g., shared mental models; (4) team behaviors, e.g., turnover or absenteeism; and (5) norms, e.g., expected behaviors. As the table proposes, effective shared decision-making may lead to positive effects at interactional levels, by which we mean, to enhanced relationships between patients and clinicians, and to positive communication and processes in clinical teams and clinical microsystems. Demonstrations that professionals are willing to inform and then respect patient preferences might lead to significant positive changes in patient experience, including satisfaction metrics, although there remains the risk that some patients will react negatively to being asked to participate in decisions and become more aware of uncertainty about treatment effects. Clinical encounters that become more difficult to conduct because patients might not want to take responsibility for participating in decisions could lead to clinicians who avoid this approach—such a consequence would need to be monitored. Longer term interactional consequences of adopting shared decision-making might be the development of a culture where careful deliberation and caution are regarded as guiding principles, where patients are coached to assess the long-term value of interventions, to question the benefits and assess their burdens. Given the evidence that teamwork failures correlate strongly with technical clinical errors, sustained collaborative deliberation with patients across teams might lead to social norms that become self-sustaining, which could be measured. Resource utilization and other performance scores may improve at group levels.
Organizational level consequences
We view a healthcare organization as one composed of multiple delivery teams or microsystems and of a size that needs management by higher order functions such as finance, human resources, and so on, led by an executive group. For an organization to accomplish shared decision-making consistently, many clinicians will need to be willing to share information with patients about alternative options with patients to support the comparisons and consider patient preferences. These tasks will need knowledge, skills, and, perhaps most importantly, a widespread shift in attitudes. As we have already noted, patient decision tools help these processes, especially if integrated into workflows and electronic record systems. However, negative consequences will also surface: organizations might resist the investment required and decline to redesign workflow to facilitate the use of patient knowledge tools or refuse to redistribute or substitute work roles. The professional shifts needed to adopt the attitudes and skills required might be too risky for organizations and could lead to workforce problems, given that the work required to confirm that shared decision-making was being provided would modify existing work patterns.
Again, achieving high-quality shared decision-making would likely lead to different scores on measures of patient experience of care which are of increasing importance as organizational benchmarks. This is especially true if the next generation of patient experience measures focus on care that has been tailored to individual preferences. Arguably, such a process could lead to positive consequences, such as fewer complaints and fewer legal challenges. Anecdotally, many organizations have invested in patient information provision, including patient decision aids, because they are convinced that well-informed patients are less likely to conduct lawsuits (personal communication, Emmi Solutions). Expecting and supporting clinicians to achieve shared decision-making could also have a positive effect on organizational morale and could over time improve lower the incidence of professional burnout by better aligning clinician’s values with the work they do. In the long term, if shared decision-making does lead to substantial changes in resource utilization—perhaps reductions in use and cost—clinical workforce composition will have to be modified. Despite the gradual shift toward value-based payment for services, some organizations, motivated by continued income derived from achieving high volumes of procedures and contacts, will see this consequence as negative. The need to consider the wider consequences of shared decision-making is reinforced by increasing acknowledgement that achieving sustainable change in healthcare delivery systems will require attention to “organizational fields” and interventions that address multiple levels of delivery systems.
Healthcare system level consequences
By system, we refer to the policies and incentives that guide the work of organizations. Predicted effects at the system level are related to the power of norms and routines—in other words, “how things are normally done,” especially if these are visible, monitored, and rewarded. The proximal effect of embedding collaborative deliberation in a normative set of operational delivery processes is the influence it might have on setting the expectations for new recruits, having to learn shared decision-making, and how to use patient decision aids. Setting such norms reinforces the predicted distal and distant effects, where greater caution and skepticism may become the expected organizational cultures in healthcare settings. This new norm would contrast with settings where the early adoption of innovative treatment is routinely viewed as positive, as a way to be at the leading edge, with little consideration of a benefit-cost consideration. Greater emphasis on supporting patients to self-manage long long-term conditions, and have realistic expectations, may improve cost-effectiveness. Systems that build, sustain, and maintain such norms are likely to have significantly lower resource utilization and reductions in malpractice costs than those that resist collaboration with patients and continue frequent contacts to generate income.
This section describes many possible consequences about which little research evidence is available. To illustrate some potential ways forward, we have outlined a selection of possible research questions; see Table 2.
이 글은 위 출처의 원문을 그 라이선스(CC BY 4.0)에 따라 옮겼습니다. 참고문헌 목록과 인용 번호, 그림 · 표 · 각주 · 수식 · 부록은 빼고 절별로 나누어 실었으며, 그 밖의 문장은 원문 그대로입니다. 원 저작자와 발행 기관은 이 화면을 만들거나 검토하지 않았고, 이곳의 해석과 낱말 풀이는 공식 번역이 아닙니다.
