The Three-Step Model

Achieving SDM is contingent on creating an environment where individuals can act independently to make choices. To create this environment, patients must be provided with information and support through the decision-making process. The three-step model (Elwyn et al., 2012) illustrates the process of moving from initial to informed preferences. In the first step, choice talk, the physician provides quality information and assesses the patient’s knowledge of options. Supporting patients’ deliberations about their options, physicians can help patients understand that reasonable options exist, correct any misinformation, and check the patient’s reaction. In the second step, option talk, patients are informed about treatment options in detail, the harms and benefits of each option, and provided with decision support. During, the last step, decision talk, patients are supported to explore what matters most to them and aided in moving toward a decision. Elwyn et al. suggest that this model should be used in conjunction with decisions support tools such as decision aids, booklets, and websites. This model is concise and easily integrated into routine clinical settings.

Despite the strengths of Elwyn et al.’s (2012) model, Stiggelbout, Pieterse, and De Haes (2015) argued that the third step, decision talk, should be split into two distinct steps. The resulting model consists of four distinct phases: The physician (1) informs the patient that a decision is to be made and that the patient’s opinion is important, (2) explains the options and the benefits and side effects of each option, (3) discusses the patient’s preferences and supports the patient in deliberation, and (4) discusses the patient’s decisional role preference. After this, a decision is made or deferred, and follow-up is planned. When applied to clinical settings, this model effectively facilitates patient-physician communication without adding to the appointment length (Stacey et al., 2017).

Collaborative Deliberation Model

Most theoretical decision frameworks focus only on the patient as an “individual”; the contribution of others outside the healthcare system is rarely considered. Elwyn and colleagues (2014) assert that the dominant view of decision making has overlooked the reality that patients turn to others for help and support. As a result, they created a more specific conceptual model, the Collaborative Deliberation Model (CD), which contains five propositions as a foundation for clinical communication processes. The first proposition takes place under constructive engagement, which occurs when two or more people explicitly engage in dialog in which interactions are characterized by curiosity, respect, and empathy. The second proposition, recognition of alternative actions, occurs when alternative courses of action, including no change or no intervention, are available and acknowledged. The third proposition, comparative learning, posits that individuals engaged in CD should compare alternative sources of action. The fourth proposition, preference construction and elicitation, occurs when interaction between individuals, consider, construct, and elicit preferences in relation to alternative course of action. The final proposition, preference integration, emerges when individuals integrate preferences in determining the subsequent courses of action.

Time

Time is the most reported perceived barrier for both patients and providers. Many patients, sensitive to the demanding workload of their providers, feel there is not enough time during clinical consultations for SDM (Fraenkel & McGraw, 2007; Frosch, May, Rendle, Tietbohl, & Elwyn, 2012). Patients reported feeling guilty about monopolizing their providers’ time and often end clinical visits short when waiting rooms are full (Bastiaens, Van Royen, Pavlic, Raposo, & Baker, 2007; Beaver et al., 2005). Similarly, providers across many different cultural and organizational contexts reported time constraint as the largest barrier for implementing SDM (Ford, Schofield, & Hope, 2003; Graham et al., 2003). However, to date there is no conclusive evidence supporting the claim that SDM takes additional time (Stacey et al., 2017). A study of medical residents found that most predictors of SDM, including physician’s gender, lost statistical significance after controlling for consultation duration, showing the important of length of time in an SBM discussion (Pellerin et al., 2011).

Relationship With Healthcare Team

Some patients believe that if they do not have a personal, trusting relationship with their provider, they cannot engage in SDM. Many patients feel uncomfortable discussing sensitive information with a new provider (Doherty & Doherty, 2005). Patients are more willing to ask questions, share personal information, and discuss concerns if they have a trusting relationship with their provider (Fraenkel & McGraw, 2007; Peek et al., 2010). However, it is not this simple, as trust can also act as a barrier to SDM. Patients who trust their providers may engage in passive behaviors, letting the professional make the decisions; patients who distrust their providers may ask additional questions and encourage SDM (Belcher, Fried, Agostini, & Tinetti, 2006). As the U.S. healthcare system moves towards integrative care across multiple providers, many patients report poor communication with and between providers, which can decrease SDM (Larsson, Sahlsten, Segesten, & Plos, 2011).

Clinical Setting

Patients report that a perceived lack of privacy, noisy environments, and feeling uncomfortable during a physical examination all act as barriers to SDM (Légaré, Ratté, Gravel, & Graham, 2008). For example, for patients suffering from hearing loss, a noisy environment may prevent participating actively in a decision consultation. Additional environmental barriers reported were poorly lit rooms and uncomfortable room temperatures.

Power Imbalance

Authoritarian or dismissive healthcare providers who dominate decision-making encounters, do not listen to or respect patients’ concerns, or use negative verbal or non-verbal behavior, impede SDM (Eldh, Ekman, & Ehnfors, 2006). Some patients perceived being a “good” patient as a passive one, who allows the provider to make the decisions on their behalf. Some patients worry that by asking questions they will be labeled as “difficult” and receive poor quality of care (Clover, Browne, McErlain, & Vandenberg, 2004).

Lack of Knowledge

Many patients feel that they have insufficient information about their condition and are unprepared to choose between options; therefore they cannot be involved in shared healthcare decisions (Beaver et al., 2005). Additionally, the use of medical terminology by healthcare providers can be a barrier to SDM. Patients often feel that providers are “talking another language” (Nordgren & Fridlund, 2001) leading patients to sometimes misinterpret or not fully comprehend the information being shared.

Providers have reported a lack of agreement with some aspects of SDM, an unfamiliarity with SDM, or that SDM is not applicable in certain clinical situations or with specific patient populations (Cohen & Britten, 2003). Some providers felt their lack of familiarly with SDM was a barrier to implementation of SDM, suggesting that interventions to train providers in SDM may be needed (see later section).

The Decisional Conflict Scale

One of the earliest scales, the Decisional Conflict Scale (DCS), was developed and validated by O’Connor (1995) in response to the dearth of instruments available to evaluate decision aids and decision-supporting interventions. The DCS was developed using items derived from the decisional conflict construct that assess uncertainty, factors contributing to uncertainty, and perceptions of effective decision making. The DCS is a 16-item self-administered questionnaire with five subscales: certainty, information, clarification of values, support or pressure from others, and the respondent’s perception of the quality of the decision process. The scale has been adapted and tested among health professionals and is widely used (Légaré et al., 2012).

Control Preference Scale

In its original form, the Control Preference Scale (CPS) was developed to measure how treatment decisions and decisional control preferences are made among people with life-threatening illnesses (Degner et al., 1997). The CPS presents patients with five alternative decisional roles and asks them to select the option which most closely reflects their perceptions of their involvement, from patient-only through shared decision making to physician-only, i.e., “I made my decision alone”, “I made my decision alone considering what my doctor said”, “I shared the decision with my doctor”, “My doctor decided considering my preferences”, and “My doctor made the decision” (Kasper, Heesen, Köpke, Fulcher, & Geiger, 2011). Although the CPS is commonly used, the patient or physician’s cultural expectations for what patients’ involvement should be might affect responses (Henrikson, Davison, & Berry, 2011).

Observing Patient Involvement in Decision Making

This observer rated instrument, abbreviated as OPTION, measures the extent to which clinicians involve patients in the decision-making process (Elwyn et al., 2003, 2005). Although many other observer instruments now exist, OPTION is still the only instrument that focuses on behaviors initiated by the healthcare provider. Nicolai et al. (2012) published a systematic review of studies using OPTION, finding psychometric and methodical limitations of the scale, including the fact the majority of studies found it to have multiple factors and that inter-rater reliability was overestimated and varied greatly across studies (Nicolai, Moshagen, Eich, & Bieber, 2012). Because of these criticisms, Elwyn, Tsulukidze, Edwards, Légaré, and Newcombe (2013) created a short version, the OPTION-5, with improved reliability and construct validity. OPTION-5 is based on a more robust conceptual model, and offers the ability to give “credit” to providers who make efforts to engage patients in decision making over multiple encounters (Barr et al., 2015; Elwyn et al., 2013). OPTION 5 is able to differentiate between various levels of patient involvement (Stubenrouch et al., 2016).

SDM-Q and SDM-Q DOC

With the increased attention on SDM, the a consortium organized by the German Ministry of Health reached the conclusion that no theory-driven, psychometrically sound self-assessment tool existed to measure the process of SDM from the patient’s perspective (Simon et al., 2006). To close this gap, two parallel instruments were developed, the Shared Decision Making Questionnaire SDM-Q, for patients and the Shared Decision Making Questionnaire for Physicians (SDM-Q-DOC). The instruments were tested in a German primary care sample of 2351 patient, yielding high internal consistency (Cronbach’s α >.90) and factorial validity (Kriston et al., 2010).

A systematic review of intervention studies found that using a shortened nine-item version of the SDM-Q (SDM-Q-9) could be used successfully to investigate the effectiveness of interventions aimed at the implementation of SDM and as a quality indicator in health services assessments (Doherr, Christalle, Kriston, Härter, & Scholl, 2017). Subsequently, the SDM-Q-9 was translated into English (Kriston et al., 2010; Scholl, Kriston, Dirmaier, Buchholz, & Härter, 2012) and tested in the U.S. with a stratified primary care sample, which confirmed its unidimensional structure and high internal consistency.

CollaboRATE

The CollaboRATE scale was developed to be a brief (three-question) measure that could easily integrate a SDM measurement into any clinical setting (Elwyn, Barr, et al., 2013). Additionally, aggregation of large numbers of responses can be used as a performance metric or feedback tool within a hospitals or clinic and to make comparisons (Barr et al., 2014). Patients rate the extent to which three core SDM dimensions are present in a specific clinical encounter (“how much effort was made to help you understand your health issue”, “how much effort was made to listen to the things that matter most to you about your health issue”, and “how much effort was made to include what matters most to you in choosing what to do next?”). CollaboRATE has been used successfully in diverse primary care settings (Forcino et al., 2017). A recent intervention, “OpenComm”, aiming to increase patients and primary care physician’s communication, found that compared with visits in the usual care clinic, patients in the “OpenComm” condition had 1.52 times higher odds of giving their primary care providers the highest possible CollaboRATE score (Tai-Seale et al., 2016).

Assumption 1: Patients Need to Have All Available Information to Make an Informed Decision

It seems logical that one can only make a decision about a situation when one is well informed. However, this assumption elevates one type of information, specifically medical risks and benefits, over any other type of information. As Leventhal and colleagues have argued in their common sense, self-regulation framework, patients often value their own experiential knowledge more than expert knowledge when making health behavior choices (Leventhal, Brissette & Leventhal, 2003). Examples range from patients who decide not to take their medication because their body doesn’t need it to the reflexive response of cancer patients who say that they would never have surgery. In these cases more information may not change a person’s mind; rather, different experiential and culturally appropriate information needs to be considered for SDM.

How much information do individuals need to make a decision? Proponents of the shared decision paradigm argue that patients should be “fully” informed. It is unclear what this really means, but what we can safely say is that there will be great variation in the amount of information that patients can process. Take the decision for prostate cancer treatment as an example. When diagnosed early, men have the choice between active surveillance, (i.e., actively and regularly monitoring the slow progress of disease), surgery, and radiation treatment (external beam radiation or radiation through implanted radioactive seeds). Each treatment option has multiple pros and cons that influence future quality of life. Patients become overwhelmed by the sheer number of factors to consider in making a decision, (mis)understanding the statistical presentation of risk and the uncertainty of outcomes. Often, patients end up asking the physician “What would you do in my place?” Whatever the circumstances, patients’ bring both their knowledge and information processing capabilities (often termed health literacy) to the situation. When these factors are not queried in a decision-making encounter and only the standard discussion of medical pros and cons is presented, the medical information becomes much less influential in the final decision.

Assumption 2: Patients Are Rational Decision Makers

Tversky’s and Kahneman (1974) seminal paper on heuristics and biases identified the pitfalls of rational decision making. Similar processes operate with medical information processing. Decisions are shaped by previous experiences, social norms, contextual factors, and cognitive biases. Biases and heuristics that have been identified are the anchoring effect, loss aversion bias, social norms bias, and action-oriented bias (Beckmann & Kuhl, 1984; Furnham, 1986; Kahneman & Tversky, 1986; Strack & Mussweiler, 1997). None of these biases or heuristics is considered in the usual presentation of information, yet they influence decisions (Chapman & Coups, 2006; Peters, Lipkus, & Diefenbach, 2006).

Cultural norms and experiences influence decisions, as well as lived or witnessed experiences. Lived experience is particularly powerful, especially when the consequences are considered painful or harmful. A man with prostate cancer might not consider surgery because a close family member who had surgery experienced persistent urinary and sexual dysfunction. This real-life experience is so powerful that it overrides “rational” arguments. In the age of social media and the Internet, lived experiences might not even be necessary any more to shape opinions; witnessed or vicarious experiences could be just as powerful in shaping opinions.

Assumption 3: Patients Make Decisions Without Emotions

Medical decisions are distressing and anxiety provoking, especially for serious or life-threatening diseases. The literature is filled with reports from patients who report that a cancer diagnosis, for example, is “like a blow to the head” and that patients are filled with an existential dread, unlike anything they have experienced before (Meyerowitz, Williams, & Gessner, 1987). Although most clinicians postpone a serious discussion of treatment options to a later date, under the assumption that patients would be less distressed, the knowledge that the disease is present is a constant source of distress. It is therefore unavoidable that important decisions are made in an emotion-laden environment.

A case in point is mammography screening among a group of women at high risk for breast cancer. In one study (Diefenbach, Miller, & Daly, 1999), women with a first degree relative diagnosed with breast cancer received genetic counseling and the recommendation to obtain a mammogram within a year. At study entry, participants were asked how worried they were about breast cancer and how likely it would be that they would be diagnosed with breast cancer in their lifetime. Mammography uptake a year later was predicted by age, education, risk perception, and worry; without worry in the model, higher risk perceptions of breast cancer were predictive of greater uptake. However with worry in the model, risk perception became non-significant, indicating that worry about cancer was the stronger predictor. Hence, the mammography decision was driven by the “hot” emotional fear of cancer and not the “cold” cognitive risk perception. Clearly, emotions influence decisions, but they are not part of any SDM model.