Shared Decision‐making in Autoimmune Neurology

Abstract

Purpose of review

In this article, the author explores the use of shared decision making (SDM) in the management of the preference-sensitive condition, neural autoantibody-mediated syndromes.

Recent findings

The field of autoimmune neurology lacks trials and often data to support therapeutic decisions. Treatment choices need to be made acutely, lacking crucial laboratory information and with uncertainty regarding treatment response and prognosis. This lack of data does not necessitate indecision in a population where delayed treatment may lead to poor outcomes. Over the past several decades, SDM has emerged as a model of communication enabling clinicians and their patients to explore current knowledge in the context of a patient's values and goals to arrive at joint decision, even when data are lacking.

Summary

SDM is a tool autoimmune neurologists should use to develop individualized treatment plans based on the patient's clinical presentation contextualized within specific values and preferences.

There has been sweeping progress in the past decade regarding our understanding of the pathogenesis and semiology of autoimmune neurologic diseases. However, with the description of new antibodies associated with an extensive range of neurologic syndromes comes the challenge of determining appropriate treatment regimens. Recent studies suggest that autoimmune encephalitis (AE) is as common as viral encephalitis.¹ However, there are no high-quality, prospective, randomized controlled trials (RCTs) for any antibody-mediated autoimmune encephalitis, even the most common, anti-NMDA receptor (NMDAR) encephalitis. Beyond logistical challenges of planning a rare disease trial, the issue is further complicated by strong anecdotally driven, dogmatic approaches to treatment that have evolved over years without prospective treatment trials. Given the array of neurologic syndromes that even a single antibody may present with, and the difficulty of determining best treatment via RCT, a single standard of treatment for patients with autoantibody-mediated neurologic diseases still remains beyond reach.

In the past 20 years, an analogous dilemma arose in the field of multiple sclerosis (MS) with the emergence of approximately 15 novel US Food and Drug Administration–approved medications. Given the rapid emergence of multiple available disease-modifying therapies (DMTs), it was not feasible to perform large RCTs comparing all DMTs. Leading MS neurologists, in recognition of the importance of respect for individual autonomy, advocated for the use of shared decision making (SDM) as a means of determining optimal treatment regimen in the setting of a complicated “benefit:risk landscape.”^2–4 As such, MS is a “preference-sensitive condition” where there are multiple treatments options with potentially similar efficacy, but with significantly different risks and benefits.^2,3,5 Accordingly, using an individualized approach, focusing on patient values with particular attention to communicating “risks, uncertainties, and options” enables neurologists to assist in decision making in the best interest of the patient despite the lack of a single consensus in the community.³

Given autoimmune neurology's emergence from the field of MS over the past decades, adopting its tool kit for autoantibody-mediated neurologic syndromes, another preference-sensitive condition, is a natural next step. Furthermore, given the lack of RCTs and federally approved therapies, SDM stands to be of great value to this branch of neuroimmunology. Last, as the next generation of autoimmune neurologists learns how to practice in a field lacking high-level evidence, SDM allows us to help make timely decisions in the face of diagnostic and therapeutic uncertainty.

Shared Decision Making

SDM is “the process by which a healthcare choice is made by the patient [or surrogate decision maker] together with one or more healthcare professionals (HCPs).”⁶ It is the “pinnacle of patient-centered care” and the mainstay regarding decision making in complex neurologic diseases, of particular utility in the setting of medical and therapeutic uncertainties.⁷

Given the rapid emergence of multiple available DMTs, it was not feasible to perform large RCTs comparing all DMTs.

Historically, SDM emerged in response to several complex issues. First, medicine moved away from physician-centered model (paternalism) and toward a patient-centered model, driven by respect for autonomy. However, this latter approach also has limits. Placing exclusive responsibility of decision making on a sick person or their family can burden rather than embolden them, and generally, people are not trained to make complex decisions in the stress of illness. The SDM model emerged “widely acknowledged as a reasonable middle ground” between paternalism and unfettered autonomy, in recognition “that patient and physician are making a common cause against illness and suffering,” leading to improved satisfaction for both parties.^8,9

Next, although many fields abound with level 1 evidence, many neurologic subspecialists depend on level 5 evidence to support therapeutic decisions.¹⁰ Divulging uncertainty can be difficult for a provider and frightening to individual who is seeking help from an expert. In addition, individuals and their families have access to online sources of varying degrees of scientific validity. SDM allows people and their HCP a venue to explore data, treatment options, and a joint opportunity to critique known information.

SDM also emerged in the context of informed consent (IC). A discussion where risks and alternatives can be discussed to ensure respect for patient autonomy further equilibrates the power differential between the doctor and the patient.¹¹ Given inherent asymmetry of information, integration of SDM into the IC process is recommended.¹¹ Furthermore, there is literature supporting an analogous approach in the management of patients who request physician aid in dying as a means of managing disagreements and ethical complexity.^12,13

Many HCPs already use SDM in their patient management. However, studies suggest that communication skills may be poor among physicians. So with the goal of further improving the communication and decision-making skills of autoimmune neurologists who manage AE and other neural autoantibody-mediated disorders, we will outline the steps of SDM demonstrating the clinical value this tool has to offer.^4,14

Step 1: Define and Explain the Problem

Before engaging in an explanation of autoantibody-mediated neurologic disorders, a neurologist must determine a person's medical literacy, interest in explanation of pathology, and desired level of engagement in decision making. People require different levels of information ranging from those who gather all available data to those who wish to remain intentionally uninformed. Notably, a person with a disease may disagree with their family on how much explanation they require and medical background does not necessarily indicate desire for detailed exploration of pathophysiology.

This initial conversation is particularly pertinent when applying SDM to AE and other autoantibody-mediated neurologic disorders, as explanation quickly wades into uncertainty. Patients and their families may ask seemingly simple questions, which remain beyond the limits of currently known information. What is known is that there are >15 unique antibodies associated with AE alone and more uncovered each year.¹⁵ Early treatment of AE is essential, and neurologists must often make therapeutic treatment recommendations with only history and limited clinical information, lacking autoantibody evaluation.¹⁶ The neurologist should discuss the nuance of the diagnosis and relay this uncertainty early to set reasonable expectations from the outset.

Regarding diagnosis, the clinical approach position paper for AE stratifies AE diagnoses into 3 categories.¹⁶ The term definite AE is reserved for when clinicians have identified specific autoantibodies in a patient with phenotypic AE.¹⁶ Probable and possible AE categories indicate that autoantibodies have not been identified, but AE is supported by clinical manifestations including history, MRI, EEG, and CSF findings such as lymphocytosis, unique oligoclonal bands, and immunoglobulin index.

Conversely, the presence of autoantibodies does not necessarily predict robust immune responsivity or pathogenesis as antibodies may represent cell surface or cytoplasmic targets. In leucine-rich glioma-inactivated 1 encephalitis, for example, antibody titer may reflect disease severity and prognosis, with higher titers associated with less vigorous response to therapy.¹⁷ Conversely, stiff-person syndrome is associated with glutamic acid decarboxylase 65-kDa isoform antibodies, but the antibodies are not clearly pathogenic, and the response to immunotherapy is variable at best.¹⁸

Last, autoimmune-mediated neurologic syndromes such as cerebellar degeneration, myelopathy, neuropathy, epilepsy, and dysautonomia do not have robust position papers, and recommendations are based on case series and expert recommendations. Overall, diagnostic discussion should be tempered, particularly if there are no autoantibodies found.

Step 2: Present Options

Treatment options for neural autoantibody-mediated syndromes depend on identified autoantibodies, disease phenotype and severity, and the person's individual goals and preferences. In AE, the sickest patients are diagnosed while intubated and sedated in intensive care. On the opposite end of the spectrum, some patients have mild symptoms and make a spontaneously full or near full recovery without pharmacologic intervention with minimal or no neurologic sequelae.¹⁹

Experts currently recommend early treatment with IV methylprednisolone, along with IV immunoglobulin or plasma exchange (PLEX) therapy for AE.^20,21 Notably, in the setting of acute psychosis associated with AE, methylprednisolone may exacerbate psychiatric symptoms, along with the complex side effect profile of steroids. However, PLEX is not always a reasonable alternative treatment option, as an agitated patient may be intolerant of prolonged use of a central venous catheter.

Unfortunately, patients with AE are often refractory to this first-line therapeutic trial, and a stepwise escalation approach is generally practiced.¹⁶ Second-line treatments include chemotherapeutic agents such as rituximab and cyclophosphamide, and in NMDAR encephalitis, second-line therapy is commonly more effective.^16,22 Recent data in NMDAR suggest that early treatment leads to improved outcomes, and some people may be better served by starting with second-line therapy, whereas others may be injured by this approach.²³ Aggressive treatment with chemotherapy is not a guarantee that a patient will improve, and despite maximal immunotherapy, some people will die of complications of AE.

Setting reasonable expectations at an early stage is essential, and aside from studies in NMDAR encephalitis discussed below, there is no way to predict treatment response.²³ People have a range of therapeutic response ranging from complete resolution, transient response with quick return to clinical nadir, or in some cases, no response at all. Unfortunately, some patients may demonstrate further clinical deterioration and die.

Initiation of primary or secondary immunotherapy MUST be undertaken with clearly outlined endpoints, i.e., predefined objective parameters defining treatment success or failure. These parameters will vary based on clinical phenotype but could range from improvement/stability in neuropsychological testing to resolution of seizures. A “time-limited trial” should be planned over a specific duration to provide interventions to give people the potential to improve but reduce “indefinite exposure to burdensome treatments.”²⁴

Although there is general consensus on management of AE, other neural autoantibody syndromes do not have robust community agreement. For these syndromes, neurologists should discuss aggressive immunomodulatory therapy, as well as conservative treatment options when warranted and/or patient preferred, including exclusively symptomatic treatment. The neurologist's recommendations should have immunologic rationale specific to the presenting pathophysiology, which should be shared with the patient. Importantly, everyone has the right to make their own choices, even if it is not the same decision a neurologist may make if they were in the patient's position.

In a contrasting situation, neurologists should not offer options outside the limits of medically or ethically appropriate treatment. Notably, 54% of low titers of nicotinic ganglionic acetylcholine receptor autoantibody (α3-AChR) are in patients with “nonautoimmune or no neurologic disorder.”²⁵ Patients with nonspecific symptoms may demand aggressive treatment posing risk without potential benefit. Given this evolving power, dynamic Virginia and Texas have passed laws to support physicians who refuse to offer medically inappropriate treatment.^26,27

Step 3: Discuss Risks and Benefits

Leaving behind ambiguity, discussing the risks and benefits of pharmacotherapy offers an opportunity to share robust data with patients. Risks of methylprednisolone, rituximab, cyclophosphamide, and mycophenolate are broadly known, and higher efficacy therapies are accompanied by increased risk. Specifically, rituximab is associated with risk of upper respiratory and urinary tract infections and progressive multifocal leukoencephalopathy and reactivation of hepatitis B.²⁸ Associated cytopenias require frequent laboratory monitoring.²⁸ Cyclophosphamide risks include myelosuppression with increased risk of infections, cardiotoxicity, pulmonary and urinary tract toxicity, secondary malignancies, liver disease, and infertility in men and women.²⁹ Mycophenolate can cause cytopenias with increased opportunistic infection risk, increased risk of lymphoma, skin cancer, and gastrointestinal upset.³⁰ Mycophenolate and cyclophosphamide also necessitate effective contraception.

However, benefits of treatment are less certain with a broad range of therapeutic response. The recently published NMDAR Encephalitis One-Year Functional Status score is the first prognostic tool autoimmune neurologists have to counsel patients on predicted functional status of patients with NMDAR encephalitis at 1 year²³

Furthermore, it is unlikely that clinicians will have autoantibody panels on serum or CSF when they must make decisions for first-line treatment in AE. All too commonly, a negative autoantibody panel returns despite clinical suspicion. In situations where there is exclusively supportive but nondiagnostic evidence for an immune-mediated disease, or where no antibody has been identified, physicians should offer realistic expectations and the range of potential therapeutic response.

Aside from AEs, in other autoantibody-mediated neurologic syndromes, similarly there can be a variety of response. The degree of familiarity with this diverse set of phenotypic expressions often requires referral of a patient to a tertiary care center where an experienced autoimmune neurologist can medically manage acute and chronic issues including immunotherapy monitoring, counseling regarding end points, prognostication, and symptomatic management. A long-term follow-up plan with a specialist is essential to establish early in treatment, and daily or weekly reevaluation of goals of care is essential, given the frequent prognostic uncertainty.

Step 4: Clarify Individual's Values, Preferences, and Goals

Decision making is inherently value-laden, and examination of values will help people rank order options, leveling the power dynamic. To arrive at a joint decision, neurologists should elicit values, goals, preferences, risk tolerance, and nonmedical circumstances like life experiences and psychosocial circumstances, either from the patient or surrogate if necessary.

Curiously, while taking a social history is a skill learned in medical school, value solicitation is not. Given the challenge and personal nature of value solicitation, neurologists need to be flexible regarding when this step is discussed in the visit. It is essential that strong rapport be established; however, patients and their families may become overwhelmed if the discussion occurs too quickly or too late in the encounter. Technical discussion can even be embedded within value solicitation where “the discussion move[s] back and forth from preferences to reasons and values to information and back.”³¹

Values are dynamic and can change when faced with mortality. Potential of severe neurologic disability can transform a conservative patient with low risk tolerance into one willing to try aggressive measures.³² Furthermore, when patients lack capacity, it is duty of their surrogate to act in accordance with the principle of substituted judgment to “try to make the decision that the patient would have made if they were able to make decisions.”³³ Surrogates should operate according to the principle of substituted judgment in value solicitation in addition to decision making.

When people are able to participate, they may make decisions differently than their surrogate might make for them or than their care partner would make for themselves. It is ethically acceptable for an individual to make a decision, like choosing aggressive therapy because their care partner feels strongly in support. A care partner is affected by a disease, especially if it limits independent activities of daily living at home. However, it is essential that individuals make decisions voluntarily outside the setting of coercion.³⁴ Asking for a one-on-one discussion to clarify that decisions are free from coercion may be indicated in sensitive situations.

The neurologist too should acknowledge their own values. More complex patients tend to be referred for evaluation in tertiary care centers, which may alter practice style and risk tolerance, as may research interests.

Step 5: Discuss the Individual's Ability and Self-efficacy

Before treatment, neurologists should evaluate the patient's support system. If considering cyclophosphamide, will they be able to adhere to the lab work? Are monthly IVIG infusions feasible, or might rituximab with a lower frequency infusion be more suitable? Even in a patient with strong self-efficacy, there are always real-world limitations. The earlier limitations are identified, the better a plan can be enacted. Similarly, the outpatient autoimmune neurologist in charge of care should be identified early; although many physicians enjoy the challenge of caring for these patients acutely, they may lack experience in the chronic care and management of this patient cohort including frequent outpatient visits in the first year. Arranging this relationship early helps to “maximize the therapeutic effects of communication… associated with improved health” and promotes “health literacy, communication, and patient and team satisfaction.”^35,36

Even in a patient with strong self-efficacy, there are always real-world limitations. The earlier limitations are identified, the better a plan can be enacted.

Step 6: Present Recommendations Based on Current Knowledge

Without a large trial, many autoantibody-mediated neurologic syndromes represent “preference-sensitive” conditions such as MS, “in which there exist a number of available treatment options of similar efficacy, with differences in risks and benefit.”^5,37 Although there are certainly widely agreed-upon circumstances where early immune therapy improves outcome, some clinicians may feel that early aggressive therapy is warranted in cases with less robust data.²³ When there is equipoise between a range of therapeutic approaches, sharing this equipoise equilibrates the power dynamic. Given this challenge, a neurologist at a tertiary care center is best suited to present their expert judgment based on the literature, experience, and ultimately contextualize options within the patient's values. Given personal value heterogeneity, there may not be a single “right” therapeutic approach.

Step 7: Clarify the Patient or Surrogate's Understanding; Step 8: Make or Explicitly Defer a Decision; and Step 9: Arrange Follow-up

The final steps in SDM emphasize the importance of clear communication. Similar to IC, the neurologist should ensure that the decision maker understands the conversation. Once confirmed, a decision can be made, which includes both the option to make a decision or to explicitly defer making one. Curiously after working through the process of SDM, 52% of individuals defer decision making to the HCP, but still want to engage in the process of deliberation.³⁸ Alternatively, an individual may make a decision a neurologist may not make for themselves, but we must respect people's autonomy. Intermittent monitoring should be offered if initial treatment is declined, and follow-up should be arranged for further management.

SDM is Not Time Intensive

Incorporating the steps of SDM may appear time consuming. However, when properly trained, patient visits do not take significantly longer, and appropriate SDM can occur in 10-minute appointments.¹⁴ Ultimately, SDM should save time for neurologists in the long run, given improved adherence to treatment plan and better outcomes.¹⁴

The dramatic presentation of these illnesses, and past anecdotal experiences with similar patients, can overshadow appropriate recognition of the currently sparse data and bias clinicians in situations requiring more equipoise than expert opinion. SDM offers neurologists a tool to quickly arrive at the best plan of care for patients with autoantibody-mediated syndromes, a preference-sensitive condition with therapeutic and prognostic uncertainty.

The therapeutic unity alters the power dynamic inherent in the doctor-patient relationship, enabling individuals and their neurologists to reach a balanced shared decision, free from the extremes of paternalism or unmitigated patient autonomy. As more data emerge in this diverse field, SDM will remain the pinnacle of patient-centered care and should continue to be the primary means of decision making as our understanding of these autoimmune neurologic diseases increases.

Appendix. Author

Study Funding

No targeted funding reported.

Disclosure

D.A. Lapides has served on an advisory board for Biohaven Pharmaceuticals. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.

TAKE-HOME POINTS

• → Other than NMDAR encephalitis, there remains prognostic uncertainty and limited data regarding management of neural autoantibody-mediated syndromes.

• → SDM represents a medium between extremes of paternalism and unconstrained patient autonomy where a neurologist guides patients through knowns and unknowns in their condition, and ultimately, a decision is made in the context of a person's specific values and goals.

• → Although SDM is an ideal tool for decision making in general, in preference-sensitive conditions such as neural autoantibody-mediated syndromes, where there are insufficient data regarding treatment, SDM offers autoimmune neurologists a means of arriving at individualized treatment plans respecting the diversity of people and the heterogeneous disease phenotypes.