Treatment preference among patients with spinal muscular atrophy (SMA): a discrete choice experiment

Spinal muscular atrophy (SMA) is a rare and debilitating genetic neuromuscular disease caused by a loss of function mutation or deletion of the survival motor neuron gene 1 (SMN1) [1‐3]. It affects one in approximately 15,000 live births [4, 5]. SMA is characterized by progressive loss of motor neurons, muscle weakness, and atrophy [6]. There are five primary types of SMA numbered from 0 (most severe) to 4 (least severe), based on age of onset and highest physical milestone achieved. The disease is a continuum from most severe to the milder phenotype form. Patients with type 0, a very rare and the most severe form of SMA, have onset of symptoms at birth and usually die beyond 6 months of age. Type 1 SMA is the most common form of SMA with onset of symptoms between 0 and 6 months. Patients with type 1 SMA present a number of functional impairments, such as severe hypotonia, progressive loss of motor function, impaired swallowing and respiratory function, and cannot sit independently during their reduced lifespan. Patients with type 2 or type 3 SMA experience disease onset before and after the age of 18 months, respectively. Those with type 2 SMA are unable to stand or walk without support, in contrast to patients with type 3 SMA who are able to stand and walk until the disease progresses. Type 4 SMA onset is in adulthood and that causes milder muscle weakness [7].

SMA is associated with direct and indirect costs of approximately $1 billion annually in the United States [8, 9]. Given the rarity of SMA, the direct and indirect costs per capita are close to $200,000 [10, 11]. The latest treatment patterns and economic study on patients with SMA found significantly higher healthcare utilization and costs than the general population [12]. Multiple studies have found substantially higher costs for patients with SMA compared to healthy populations [10, 11, 13]. In addition to impairment on quality of life and social participation among patients with SMA, SMA was associated with reduced caregiver quality of life, which was related to patient burden, impaired psychosocial well-being, and loss of caregiver’s productivity [14‐16].

Despite the progress in SMA treatment with three important landmark treatments (approved by the Food and Drug Administration [FDA]), the disease burden and related unmet need remains significant for many patients with SMA and their caregivers [5, 17]. Spinraza® (nusinersen) is a disease-modifying treatment approved for all patients with SMA and is administered via intrathecal (IT) injection every 4 months after a series of loading doses [18]. While repeated intrathecal injections are the standard approach to achieve adequate intrathecal distribution and targeting of the central nervous system (CNS), alternatives such as intrathecal catheter systems are being explored [19, 20]. Nusinersen is a survival motor neuron-2 (SMN2) directed antisense oligonucleotide that enhances production of full-length survival motor neuron (SMN) protein and slows the progression of the disease. Zolgensma® (onasemnogene abeparvovec-xioi) is a gene replacement therapy that uses a viral vector to deliver a functional copy of the SMN1 gene. Onasemnogene abeparvovec-xioi is administered as a one-time intravenous (IV) infusion and is only approved for the treatment of children less than 2 years of age because of current limitations of dosing (i.e. viral titers and increased likelihood of immune response) and the fact that this drug has only been tested for this age group [17, 21, 22]. The most recent treatment approved by the FDA is Evrysdi® (risdiplam), an orally administered, SMN2 splicing modifier for patients 2 months of age and older with SMA [23, 24]. The drug increases exon 7 inclusion and thus full-length SMN protein production from the SMN2 gene [23]. The SMA treatment landscape continues to evolve as there are several other medications in development.

The unique health status and life experience of each patient and family affected by SMA may influence how they perceive meaningful changes with respect to desired benefits or avoided risks from a therapeutic agent. With multiple treatments available, both healthcare professionals, as well as patients and caregivers, have many considerations to consider when deciding on the right treatment choices [25‐28]. A 2007 consensus statement, issued by the International Standard of Care Committee for SMA, to address the 5 priority care areas for SMA, found that for those affected by type 1 SMA, high priority components of meaningful change relate to immediate concerns with breathing, feeding, swallowing, and the ability to communicate [29]. For patients with types 2 and 3 SMA, priorities include disease stabilization, independence and the ability to perform basic personal tasks, and reducing muscle fatigue [29]. In most cases, patients and caregivers perceive benefit in maintenance of current abilities and avoiding decline in function. They indicate that even small improvements in functional abilities would be meaningful, especially if these contribute to the increase in ability to independently perform activities of daily living [29].

Despite the burden of SMA, little focus has been given to understand how patients and caregivers weigh the possible risks associated with therapies against potential benefits associated with those therapies [28]. Patients’ perceptions of benefit-risk are essential to informing the drug approval process and the context in which potential therapies are evaluated. In light of three current SMA treatments and more treatment options on the horizon, a unique window of opportunity arises to assess and characterize preferences for patients and their caregivers on SMA treatment characteristics and attributes. Therefore, the primary objectives of this study were to quantify benefit-risk trade-off preferences and to estimate the relative importance of hypothetical treatment attributes. This study addresses a significant gap in the evidence base for SMA, particularly given patient and caregiver perceptions of benefit-risk tradeoffs that inform the patient-centered value framework for the innovative therapies. Further, this study has the ability to provide evidence to support the need for informed joint treatment decision making for SMA patients by educating providers to understand patient and family preferences, and informing payers about the unique role each therapy may bring; thus, allowing equitable benefit design so patients have access to treatments that work best for them.

There were 101 respondents (65 self-reported patients and 36 caregivers) included in the analysis. Table 2 displays the clinical and demographic characteristics of both patients and caregivers. Patient ages ranged from less than 1 year old to 67 years old, 64% were female, and 81% were white. The household income across respondents varied with 21% having less than $25,000 and 22% of respondents having $100,000 or more.

Table 3 presents the clinical and disease characteristics for SMA patients. In total, there were 21 patients with type 1 SMA, 48 patients with type 2 SMA, 29 patients with type 3 SMA, and 1 patient with type 4. A majority of patients (68%) had their first symptoms before they were 18 months old. 58% of patients were on nusinersen at the time of survey completion. Almost half of the study respondents reported motor function without assistance (42%) and 84% reported breathing function without mechanical support.

Figure 2 and Appendix Table 4 present the conditional logit model estimates for the SMA treatment preferences. Overall, better improvement in motor and breathing function was highly valued (regression coefficient (RC): 0.65, 95% confidence interval [CI]: 0.47–0.83 and RC: 0.79, 95% CI: 0.60–0.98, respectively); and oral medication and one-time infusion were strongly preferred over repeated IT injections (RC: 0.80, 95% CI: 0.60–0.98 and RC: 0.51, 95% CI: 0.30–0.73, respectively). Patients least preferred an age-restricted label/approved use (≤ 2 years of age) and serious/life threatening risk (RC: − 1.28, 95% CI: − 1.47 to − 1.09 and RC: − 1.83, 95% CI: − 2.11 to − 1.60, respectively).

In the overall survey sample, when changing the route of admission from repeated IT injections to one-time IV injection, the results yield approximately 0.78 (0.51/0.65) times as much utility as increasing efficacy from stabilization to better improvement in motor function, signifying a preference for higher efficacy. On the other hand, changing route of admission from repeated IT injections to oral daily medication yields approximately 1.2 (0.80/0.65) times as much utility as increasing efficacy from stabilization to better improvement in motor function. All of the findings suggested a much greater preference weight was given to the oral daily administration, rather than IV injection. Likewise, reducing side effect severity from ‘moderate risk’ or ‘serious, life-threatening’ to ‘mild risk’ yields approximately 1.2 (0.76/0.65) and 2.9 (1.86/0.65) times as much utility, respectively, as increasing efficacy from stabilization to better improvement in motor function, signaling a stronger preference for risk minimization over additional efficacy gain.

The cross-attribute results related to breathing, such as preference on oral administration and risk aversion were similar to those of motor function. Changing the route of admission from repeated IT injections to one-time IV injection, the results yield approximately 0.65 (0.51/0.79) times as much utility as increasing efficacy from stabilization to better improvement in breathing function. Changing route of admission from repeated IT injections to oral daily medication yields approximately 1.0 (0.80/0.79) times as much utility as increasing efficacy from stabilization to better improvement in breathing function. Moreover, when looking at reducing potential side effect severity from ‘moderate risk’ or ‘serious, life-threatening’ to ‘mild risk’ yields approximately 0.96 (0.76/0.79) and 2.4 (1.86/0.79) times as much utility, respectively, as increasing efficacy from stabilization to better improvement in breathing function.

Figure 3 and Appendix Table 5 present the conditional logit model estimates for SMA treatment preferences by multiple sub-groups: (1) Types of SMA—type 1, type 2, and type 3; (2) Age groups— ≤ 2 years, 3–17 years, and ≥ 18 years; (3) Gene Therapy—Yes/No, (4) nusinersen—Yes/No; (5) Spinal surgery—Yes/No. Regardless of subgroup, all groups preferred improvement in both motor and breathing function than stabilization of these functions, all else being equal. Improvement in breathing function is valued more than in motor function for patients with type 1 and 2 SMA whereas improvement in motor function is more important for patients with type 3 SMA. When evaluating who the drug is approved for, all subgroups preferred drugs that were approved for adults and children. However, this was no longer significant when examining the subgroups of children (e.g., age ≤ 2 years old, and 3–17 years) and those who were on gene therapy, as caregivers did not have a preference as to who the drug was approved for. When evaluating the route and the frequency of treatments, all subgroups preferred the oral daily medication. Interestingly, those who were not currently under nusinersen treatment, were less likely to prefer the one-time IV infusion rather than the IT spinal injection that is done 3–6 times per year (RC: − 0.65, p value < 0.05). A related point to consider is that a stronger preference for the oral daily medication over repeated IT spinal injection was found when examining the subgroup of patients who had previous spinal surgeries than in the overall sample. The preference for oral daily is independent of SMA treatments patients were receiving at the time of the survey (except for patients not receiving nusinersen treatment).

The treatment landscape for SMA has undergone tremendous changes with three FDA approved therapies becoming available in the last few years. It is therefore critically important to understand how patients and caregivers view treatment choices. This study is the first-ever discrete choice experiment to characterize decision-making and benefit-risk trade-offs for SMA treatment characteristics. These findings can inform the perceptions and preferences of SMA patients and their caregivers to stakeholders including regulators, patients, patient advocacy group such as CureSMA, and providers.

Across patient-centered treatment attributes to address the variety and severity of impairments and disabilities that accompany SMA, findings from this study identified the overall preferred treatment attributes including improvement in motor function, breathing function, oral administration, broad indication without age limits, and minimal risk. Overall, the preference on treatment characteristics were quite robust across demographic groups and clinical subgroups. Respondents’ age, sex, SMA type and associated characteristics (e.g., age of diagnosis ≤ 18 months), did not show a difference in direction of conditional logit estimates for patient preference.

The route of administration is an important treatment attribute because of the challenges of IT administration for SMA patients, particularly for those with severe scoliosis and spinal surgeries. The challenge of IT treatment may be further complicated by its treatment burden (travel, time from school/work) to receive IT injection among adults which may provide reasoning as to why oral route of administration was preferred. It is surprising that respondents expressed a stronger preference for oral daily administration than the one-time IV infusion except for those who were treated with gene replacement therapy. One possible explanation might be that although all treatment profiles are hypothetical in this experimental design, patients and caregivers would have naturally associated treatment attributes to actual approved treatments, as such one-time treatment may have suggested gene replacement therapy for some patients, which may have influenced treatment preference. Moreover, broad indication is strongly preferred for all patients except those ≤ 2 years of age, which currently have an effective gene replacement therapy available for this age group.

In using this DCE methodology, it is important to evaluate the preferences in the context of other features, examining the relative size of those preferences compared to other preferences. In doing so, the cross-attribute results found that for most patients (e.g., very young patients, adult patients, or patients with spinal surgeries), there may be willingness to trade off higher efficacy for less invasive, less risk, and more convenient route and frequency of treatment. When reducing side effect severity from ‘moderate risk’ to ‘mild risk’ the utility was 20% more than utility of moving from stabilization to better improvement in motor function. When reducing side effect severity from ‘serious, life-threatening’ to ‘mild risk’, this ratio was more than 5 times greater yielding 2.9 (1.86/0.65) times as much utility, as increasing efficacy from stabilization to improvement, suggesting a stronger preference for risk minimization over additional efficacy gain.

Likewise, the utility of changing from repeated IT injections to oral daily medication was 20% more than the utility of moving from stabilization to improvement; for adults, this ratio was about 50% more. These findings are consistent with previous research that has found stabilization in treatment to be highly meaningful [39], and although improvement in treatment is obviously preferred if given in comparison alone, patients would prefer to maintain stabilization if that means other treatment attributes would be prioritized (i.e. route of administration, side-effect severity, etc.). In a European SMA survey conducted by Rouault et al. (2017), the authors found that 96.5% of respondents believed that “stabilization of their current clinical state through treatment would represent progress” [39]. These findings demonstrate that SMA patients and caregivers deem stabilization as a very meaningful marker for their quality of life. This should be taken into consideration by clinicians and drug manufacturers, when developing new drugs and understanding benefits and risks associated with such treatments.

A previous study assessing patients’ perceptions of benefit-risk decision-making for SMA therapies found no strong correlations between risk tolerance and patient/disease attributes [40]. However, this was before the approval of any therapeutic options. Now that new treatment options are available to patients, the benefit-risk decision making we observed in this study found that patients had a lower willingness for high risk therapies even if the benefit trade-off was higher.

Evaluating patient preferences in our study should be interpreted in the context of two innovative treatments used by some of the study respondents as two important subgroups (i.e. onasemnogene abeparvovec-xioi and nusinersen), although the use of gene replacement therapy in our study only represents a small sample size. A comparison on the conditional logit estimates found that the preferences between nusinersen users and non-users were quite similar, except for the strong preference for oral daily administration among those receiving nusinersen (repeated IT spinal injection). The Pacione et al. (2019) qualitative study on the perspectives of nusinersen treatment, characterized the patients/caregivers with SMA who did not want or were unsure about nusinersen and found that their decision about pursuing nusinersen treatment was quite nuanced, challenging and context-specific [41].

It is important to note that all treatments presented to patients and caregivers in this study are hypothetical and were automatically generated by computers. As such, one is supposed to not associate the preference (or lack of) for certain treatment attributes to actually approved treatments and treatments currently being developed. We recognized that it may be inevitable that a distinct treatment attribute may be linked to a specific treatment and/or patient experience with such a treatment, which may have influenced treatment choices. For example, the actual 10 gene-replacement therapy users did not appear to prefer the oral daily administration, and instead had a strong preference for the one-time IV administration, which may suggest the experience of one-time IV gene transfer is well tolerated and could be appealing for patients who are eligible.

With new treatments potentially turning a life-threatening disease into a chronic disease, clinicians may need to conform to a patient-centered care model that prioritizes a patient’s/family’s values, goals, and knowledge as well as patient’s self-awareness. The attention to the values and priorities from clinicians may create a shared decision making between patients and providers.

There were several limitations in the study. First, the survey sample was drawn from a SMA patient organization membership database, representing a highly informed and engaged patient population, and therefore the results may not represent all SMA patients in the United States. In particular, our sample only had one patient with type 4 SMA. Second, SMA type, motor function, and treatment status were self- or caregiver-reported. No attempts were undertaken to verify such information. Third, although affordability was a concern for many patients and caregivers, patient out-of-pocket cost was not included in the treatment profile due to administration burden and sample size limitation. Furthermore, a number of factors beyond the efficacy and side effects of treatment may influence perception and preferences on treatment characteristics among patients with SMA including disease knowledge, transportation needs, and functional status as well as insurance approval or financial assistance for very expensive SMA treatments. These factors may also interact with each other with large scope of complexity and difficulty involved in the treatment decisions for patients with SMA.

In regard to availability of actual treatments on the market, preference of one attribute may be confounded by related attributes based on actual, approved SMA treatments on the market (e.g. association of one-time infusion with gene therapy). Additionally, treatments were not explicitly described during the interview process in terms of scientific facts of effects and adverse effects of these treatments. Information about the purpose of the survey and directions for completing the survey were provided but information about specific SMA treatments and their effects were not provided. Patients only made responses based on their own personal knowledge, which limited the chance of interviewer bias. This is because the DCE study design was hypothetical and the purpose of this study was not to explicitly describe SMA treatments but rather common treatment attributes and characteristics. Thus, allowing us to understand a patient’s desired treatment in a hypothetical scenario to provide insight as to whether patients would be interested or prefer an option over the others if it was available.

Lastly, the hypothetical treatment attributes and levels may over-simplify any treatment in the real world; treatment decision has to be made in clinical context and tailored to patient specific needs. Nonetheless, these hypothetical scenarios provide insight into patient and caregiver preferences that can inform future drug development and drive the implications for future research of this highly rare, complicated, and expensive disease.

Improvement in motor function, breathing function, broad indication, oral daily administration, and minimal risk are preferred treatment attributes among all SMA patients. Cross-attributes trade-off decision suggested a lower willingness for a high-risk therapy despite additional efficacy gain. For some patients, there may be willingness to trade off additional gains in efficacy for a change in route of administration from repeated IT administration to oral medication. Patient preference and clinical context should be an integral part of treatment decision process and tailored to individual patient needs.