Chronic Disease Decision Making and ‘What Matters Most’

Abstract

The increasing use of the question, “What matters most to you?” is a welcome development in the effort to provide patient-centered care. However, it is difficult for clinicians to translate answers to this question into treatment plans for chronic conditions, including recognizing when to consider options other than clinical practice guideline (CPG) – directed therapy. Goals elicitation is most helpful when a patient has different treatment options with clearly identifiable trade offs. In the face of trade offs, goals elicitation helps patients to prioritize among potentially competing outcomes. While decision aids (DAs) focus on trade offs by delineating options and outcomes, the robust outcome data necessary to create DAs for older patients with multimorbidity is often lacking and even mild cognitive impairment makes the use of DAs difficult. The challenge for providing chronic disease care to older patients who are at risk for adverse events from CPG-directed therapy because of multimorbidity and/or frailty is to organize the complexity of individual combinations of diseases, conditions, and syndromes into common sets of trade offs and to identify those goals or priorities that will directly inform a plan of care.

Walking around the inpatient wards and outpatient clinics of healthcare facilities, one is more likely than ever to hear providers talking about “goals of care discussions” and finding out what matters most to patients. This is a welcome development. It follows several decades of scholarship on shared decision-making (SDM)¹ and increasing awareness of the narrowed net therapeutic benefit that patients with serious illness, multimorbidity, and frailty experience from clinical practice guideline (CPG) recommendations.² Because many of these patients are of older age, identifying what “matters most” has become one of the Geriatric “5Ms”: the core competencies in geriatric medicine.³ The fundamental principle is that understanding a patient’s goals can help to determine an overall treatment approach and inform multiple decisions.

Limitations in Common Approaches to Goals Elicitation

There is no consensus about how best to elicit goals. The most commonly employed approaches may not accomplish the objective of ensuring that patients’ values inform clinical decisions. Consider Mrs. G, an 80-year-old woman with chronic obstructive pulmonary disease (COPD), heart failure, atrial fibrillation, and a history of gastrointestinal bleeding whose recent unintentional weight loss, slowed pace of walking, and fatigue indicate that she is frail. If she is admitted to the hospital, the healthcare team may conclude it has conducted a goals of care discussion by establishing a “code status” for Mrs. G — a determination of the invasive interventions, including cardiopulmonary resuscitation or intubation, she would or would not want to undergo.⁴ However, asking patients about desired interventions does not capture meaningful aspects of what matters most to them. Patients’ decisions about treatment are based to a much greater extent on the outcome of the intervention rather than the intervention itself.⁵ Patients who would want to be intubated for a reversible episode of heart failure if they could return to their prior functional status would not want to be intubated for worsening COPD if this resulted in severe functional disability. Goals elicitation, therefore, needs to be outcome- rather than intervention-based.

Asking about What Matters Most for Goals Elicitation

One outcomes-based approach that is quickly gaining in popularity involves an open-ended assessment of patients’ valued outcomes. Rather than starting with healthcare decisions, it involves asking, “What matters most to you,” and having the patient identify the specific goals he or she would like to achieve.⁶ Commonly, individuals respond with statements about being able to take care of oneself and participate in meaningful activities. This approach has great theoretical appeal. Individuals are encouraged to think broadly about what matters to them. Beginning with values and goals rather than with treatment options can improve decision making by not pre-specifying what these options might be.⁷

Suppose Mrs. G says that her goal is to be able to walk to the mailbox independently. Knowing this will inform her serious illness care. Her physicians will want to counsel her about limiting aggressive interventions that can extend short-term life expectancy but at the expense of declines in function.⁸ It is less clear what her doctor will do with this information during a routine outpatient visit for management of her chronic conditions. This approach to the elicitation of goals will most likely not result in any changes to the CPG-directed care she is receiving for several reasons.

First, there are no guidelines or any other forms of instruction for how to map goals onto treatment options. Intervention studies in chronic illness rarely move beyond disease-specific outcomes to outcomes that cut across individual disease, such as function. As a result, clinicians do not have the information they need to create individualized care plans based on goals. For the many patients who will, in an open-ended elicitation process, identify a specific functional goal such as caring for a relative,⁶ there is no guidance to physicians on how to create a care plan to achieve that goal and no data to support choosing something other than CPG-directed care. In many circumstances, traditional CPG-directed care may in fact be the best plan for meeting goals. For example, CPGs provide guidance to clinicians on how to achieve relief of symptoms, which are a frequent impediment to goal attainment. If Mrs. G’s ability to walk to the mailbox is limited by shortness of breath, the clinician will want to use CPGs to optimize treatment of her heart and lung disease. Second, for a number of conditions, the same factors that place patients at increased risk from the disease may also place them at increased risk from the treatment. Therefore both disease and treatment are associated with adverse outcomes that prevent attainment of goals. Mrs. G’s age and bleeding history place her at increased risk of bleeding from anticoagulation for treatment of her atrial fibrillation. While she is at risk of functional decline resulting from hospitalization for a bleed, she is also at elevated risk of stroke and functional decline without anticoagulation.⁹ Third, there is accumulating evidence of an absence of benefit for some common interventions among vulnerable older persons, such as screening for breast or colorectal cancer screening in patients with limited life expectancy. Use of goals elicitation is frequently recommended as the approach to stopping screening¹⁰ as well as changing other established practices, such as easing glucose control among older patients with diabetes mellitus.¹¹ However, when patients do not obtain benefit, a change in practice is appropriate regardless of goals.

A recent paper demonstrates the encouraging finding of an increased likelihood of discussion or decision making concerning patients’ health priorities with an open-ended approach to goals elicitation. However, there were only modest changes in ambulatory health care activities, without clear findings regarding the consideration and/or use of treatment options other than CPG – directed therapies.¹² These findings suggest that further efforts to align care with patients’ outcome goals may benefit from identifying those patients and decisions for whom the explication of goals will inform a choice among alternative approaches to care.

Trade Offs as a Central Component of Decision Making

The distinguishing feature of decisions for which a knowledge of goals is necessary to identify the optimal approach to care is the presence of clearly identified trade offs among alternatives. Trade offs occur when the effort to achieve a desired outcome comes with the potential for or an actual undesired outcome. In the face of trade offs, individuals need to prioritize among potentially competing goals. In other words, if all goals cannot be obtained simultaneously, the optimal treatment option is the one with the greatest likelihood of achieving the patient’s highest priority goal.

Use of Decision Aids to Facilitate Decision Making

What if, rather than beginning with an open-ended elicitation of goals, decision making in chronic illness begins by identifying those patients who are facing decisions that have clearly identifiable trade offs and elicits goals in a focused way to inform those trade offs? This is the approach employed by decision aids, which contain information about the available treatment options, the likelihood of different outcomes associated with each one, and a form of values elicitation as they relate to these options. Decision aids bring trade offs into sharp focus, which can help individuals prioritize among a finite subset of possible outcomes.

Decision aids as currently utilized have limitations for the care of vulnerable older persons. Designed to address single decisions, the development and use of multiple aids would be required to inform a plan of care for patients with multiple conditions. However, it has been difficult to achieve the adoption of even single decision aids into clinical practice, with a recent study demonstrating that only a minority of aids that are shown to be efficacious in randomiozed controlled trials (RCTs) are used in clinical practice following trial completion.¹³ The most successful implementation strategy appears to be when the decision can be anticipated ahead of time and the patient works through the aid outside of the clinical encounter.¹⁴ Such aids are best suited for decisions with robust data regarding probable outcomes. In contrast, even among those decisions for which trade offs can be clearly identified for vulnerable older persons, the data are frequently limited if not altogether absent. As an example, the routine exclusion of the oldest old and patients with multimorbidity from RCTs has led to debate about the balance between benefits and harms of preventative therapies with otherwise strong evidence of efficacy, such as statins, in these groups.¹⁵ Decision making in the absence of clear data is highly complex and not well suited for patients to work through without clinician guidance.¹⁶ Finally, most decision aids utilize a formal assessment of preferences and values based on deliberate and often complex methods. Even mild cognitive impairment has profound effects on the ability to engage in this kind of thinking.¹⁷ Perhaps even more importantly, all patients bring a wide variety of beliefs, life experiences, and other considerations to medical decisions, and this is even more pronounced among older persons, who have frequently experienced their own illnesses and the illnesses of loved ones. It is challenging for structured preference assessment tools to capture all of these considerations and help both patients and clinicians to understand how they may influence patients’ decision making.

Toward a New Approach for Vulnerable Older Patients

We need to develop an approach to SDM that meets the particular needs of vulnerable older patients and builds on the strengths of each of the approaches described above. Centering the approach on the identification and exploration of trade offs will emphasize to both patients and physicians that there are options in the management of chronic disease and that CPG-directed therapy may not provide the best fit with patients’ priorities. Maintaining the notion that an understanding of patients’ goals can help to shape an overall plan of care will move the approach away from thinking about every step in chronic disease management as a separate decision. The key challenge is reducing the complexity of the many individual decisions faced by different patients with different sets of chronic diseases, conditions, and symptoms into meaningful categories that can help to guide decision making.

While this is a daunting task, there are two potential ways forward. One is to identify “classes” of decisions that share a common set of trade offs. For example, among older persons with multiple chronic conditions, medications prescribed for primary or secondary prevention have in common the trade offs between future reductions in risk and current potential for direct adverse effects, complex medication regimens, and the risks of polypharmacy. Understanding how a patient values current versus future health risks can help to inform an overall approach to the preventative medications. There are likely to be other classes, the identification of which will require a systematic evaluation to identify alterations in benefits and harms of CPG-directed therapies among vulnerable older patients. It will also require expert review and clinical judgment to identify common trade offs and optimal approaches to eliciting goals in a manner that informs prioritization of outcomes within trade offs.

A second strategy is to identify those priorities that can directly map onto and inform a plan of care. As an example, patients vary in their beliefs regarding the efficacy and safety of medications, regardless of their individual conditions, and these beliefs are associated with medication adherence.¹⁸ For patients who are strongly against taking medications, the provision of information about outcomes and trade offs is unlikely to change their minds.¹⁹ This is illustrative of the concept of “protected values,” for which individuals resist any sort of trade off.²⁰ Patients may hold other protected values relevant to medical decision making²¹ that can inform multiple decisions.

Bringing questions about what matters most to patients into the clinical conversation has been a major step forward in the care provided to vulnerable older persons. The next step in the evolution of patient-centered decision making is to find the best ways of organizing the clinical complexity of vulnerable older patients so that it is possible to translate “matters most” into a care plan for the management of chronic illness.