Preferences for Healthcare Delivery in Amyotrophic Lateral Sclerosis (ALS): A Survey of Patients and Caregivers in the United States

Abstract

Background and Objectives

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive and fatal neurological disease that leads to death within 2–5 years of diagnosis for more than 80% of people living with ALS (PLWALS). The American Academy of Neurology (AAN) developed practice parameters—general principles to guide clinicians in managing ALS—encouraging multidisciplinary care (MDC) but does not recommend specific healthcare delivery models. Three healthcare delivery models have evolved: a traditional model, a triage model, and a non-triage model. This study aims to describe preferences for and satisfaction with various models, among PLWALS and their caregivers (CALS), along with their perceptions of how their care aligns with AAN guidelines.

Methods

A cross-sectional observational study utilizing a web-based survey was distributed via email to PLWALS and CALS. Three multi-assessment questionnaires were developed and tailored for PLWALS, CALS, and former CALS. Best-worst scaling (object case) data were analyzed using a best-minus-worst approach and descriptive statistics were calculated from means, t-tests and chi-square.

Results

The combined sample included 378 respondents: 254 PLWALS (67.20%) and 124 CALS (32.80%; composed of 79 current caregivers [20.90%] and 45 former caregivers [11.90%]). The mean respondent age was 61.09 years (SD 11.1). The majority of the sample was white (92.86%), insured by Medicare (61.11%), and married/partnered (79.10%). Respondents preferred a non-triage model the most and a traditional model the least; 88.20% (CI: 84.92–91.49) were extremely likely to choose a non-triage model if given the choice and 83.12% (CI: 79.29–86.92) of respondents ranked non-triage as most preferred. A traditional model was ranked as the least preferred model in 75.28% (CI: 70.78–79.78) of respondents. The most important factors driving respondent preferences were ALS expertise and team-based care. Overall, respondents are satisfied with their care teams. PLWALS utilizing non-triage MDC models reported more adherence to quality care measures compared with those utilizing triage and traditional models.

Discussion

Respondent preference for non-triage models is consistent with the importance they place on the features of non-triage models. However, these findings should be understood in the context of our sample in which a large majority of respondents were receiving care via a non-triage model. To ensure ALS care delivery is patient-centered, practice parameters that aim to guide clinicians in managing ALS should provide more guidance to MDCs to deliver care aligned with patient preferences and values. Efforts should focus on sustainable financial models that can better facilitate non-triage models of care.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40271-026-00807-4.

Key Points

People living with ALS (PLWALS) and their caregivers (CALS) prefer non-triage multidisciplinary care (MDC).

The preference for non-triage MDC is likely driven by the importance placed on team-based care and ALS expertise.

ALS practice parameters, which are principles of clinical management, should provide additional guidance on delivering care to align with PLWALS and CALS preferences and values.

Decision makers should consider policies that help make non-triage models financially sustainable.

Background and Objectives

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive and fatal neurological disease due to motor neuron degeneration. It leads to death within 2–5 years of diagnosis for more than 80% of people living with ALS (PLWALS). There are 24,821–31,843 PLWALS in the United States (US) and the prevalence of ALS is 5.5–7.7 per 100,000. ALS symptoms include stiffness, twitching, weakening and wasting of muscles. Moreover, 15% of PLWALS in the US have frontotemporal dementia and 20–50% of PLWALS have cognitive disfunction [1, 2]. While there are some medications that slow disease progression and lengthen survival, their effect is modest, and the mainstay of care is supportive and directed at reducing disease burden and improving quality of life.

The American Academy of Neurology (AAN) developed practice parameters, which are consensus-based general principles of the clinical management of ALS in the US. The practice parameters encourage multidisciplinary care (MDC) plans, treatment for respiratory/nutritional dysfunction, use of disease-modifying agents, and plans for a smooth transition to palliative care [3, 4]. Importantly, these practice parameters recommend MDC, but do not recommend specific healthcare delivery models. AAN also developed quality measures to address supportive care, patient safety, planning for complex needs, and use of a patient-centered approach to proactive decision making [5, 6]. While these measures address the value of MDC, they focus on MDC plans that intend to facilitate quality in both solo and multidisciplinary practice environments. MDC clinics outperform non-MDC providers on survival, symptom management, and quality of life [7–12]. Thus, MDC is accepted by most as an important advancement in ALS therapy and a cornerstone of ALS care [3, 13].

MDC is a key quality measure identified by AAN, but style of delivery is not specified. As a result, depending on finances and availability, several healthcare delivery models have evolved in the US: a traditional model, a triage model, and a non-triage model. With the traditional model, the PLWALS have a neurologist visit at which time the neurologist determines what other disciplines are required and refers the PLWALS to other specialist providers, often outside of the practice. In the triage model the physician sees the PLWALS and determines which ancillary health services need to be provided to each patient on a given day, and those services are provided within the practice during the same visit. The non-triage model is a common care model in which all members of the MDC team examine the PLWALS each visit. A recent survey of 63 clinic directors in the US found that about three quarters of clinics use non-triage and about one quarter of clinics use a triage model [14].

While these models may all technically constitute ‘multidisciplinary care,’ the lack of standardization presents challenges. Clinics in the US that want to provide a truly comprehensive non-triage model face significant financial challenges [14]. Clinicians practicing in a traditional model may have greater difficulty achieving the objectives of MDC because the barriers to coordination are greater. From the PLWALS perspective, each model has trade-offs and offers quite different experiences. For example, in a non-triage model, clinic visits are lengthier but there is no need for additional burdensome trips to other specialists as in the traditional model. All these challenges associated with the lack of standardization in MDC are likely to be exacerbated in the next 5 years as ALS prevalence in the US is anticipated to rise by more than 10% [15].

The primary goal of this study was to describe preferences for the features of different types of MDC care models in the US. The inclusion of the perspectives of informal caregivers of PLWALS (CALS) in our study was important because CALS experience a high burden of care, which increases as the disease inevitably progresses [16]. In a recent survey, 68% of CALS in the US spent more than 30 hours per week caregiving [17]. In the same survey, CALS also reported that home visits from various healthcare disciplines (e.g., nurses, occupational therapists, or physical therapists) are their highest unmet need. This may indicate how system-level factors, such as how care is delivered, may contribute to caregiver burden and provide an opportunity for intervention to reduce caregiver burden. While some studies have explored the relationship between person-level factors and caregiver burden [17–20], none have explored how care delivery impacts CALS, despite that, at a minimum, CALS experience care delivery models alongside PLWALS, and for most, CALS are an essential part of the MDC team.

To understand the perspective of PLWALS and CALS, we utilized stated preference measurement as a central methodology. Stated preference measures are a survey-based preference elicitation technique that requires respondents to state their choices about hypothetical scenarios [21]. From those choices, researchers can infer preferences. Specifically, we used best-worst scaling (BWS), which is a choice-based preference elicitation method used to elicit respondent's relative prioritization for distinct items [22]. It is an application of consumer theory in which utility is derived from a particular bundle of characteristics rather than from a good itself [23]. The object case format used in this study asks respondents to identify the ‘best’ and ‘worst’ objects from a list, representing the extreme ends of a latent, subjective continuum. Although first proposed in the early 1990s [24], it has become an increasingly popular alternative to rating scales in healthcare [25, 26].

Significance

Given the various models, it is important to determine PLWALS’ and CALS’ preferences for care delivery models as this will influence adherence and trust and reduce burden of healthcare. The goal of this study is to describe PLWALS’ and CALS’ preferences for alternative care models based on how they value the features of different types of care models in the US. We also aim to measure satisfaction with their current care and adherence to care guidelines from the perspective of PLWALS and CALS.

Methods

We designed a cross-sectional observational study utilizing a web-based survey distributed via email.

Recruitment and Eligibility Criteria

Invitations to participate were emailed to people who were registered with three national organizations. The Center for Disease Control (CDC) National ALS registry is a program to collect, manage, and analyze data about PLWALS who choose to participate and submit their own information. Everything ALS is a patient-focused national advocacy non-profit organization that offers an open-data platform for direct engagement with PLWALS and CALS. The Northeast ALS (NEALS) Consortium PEACe Committee is a US-based committee dedicated to education and advocacy to empower PLWALS to have a greater understanding and a greater role in research. This committee has global reach but invitations were sent to a US-based distribution list. The three distribution lists include 11,295, 7000, and 300 emails, respectively. Because the degree of overlap between lists is unknown, the number of unique email addresses can be estimated as between 11,295 and 18,595.

A convenience sample of PLWALS and CALS were recruited as individuals, as opposed to dyads. Data were collected from November 5, 2022 to June 6, 2023. To maximize the size of the sample, the survey closed after an extended period without additional enrollments. Eligibility criteria included English-reading PLWALS and CALS or former CALS of PLWALS (if caregiving was provided within the previous 5 years). Eligibility screening was conducted via REDCap survey questions to determine participant role. ALS diagnosis (or presumed diagnosis) was also self-reported as part of eligibility screening. Respondents were encouraged to take breaks as needed and were not required to complete the survey in a single sitting. No questions were mandatory; respondents could skip any question they did not want to answer. Respondents were welcome to have assistance in completing the survey if required and were not compensated for their time.

Survey

PLWALS Version

We developed a multi-assessment questionnaire (see electronic supplementary material [ESM]). The final survey instrument captured preferences for care model characteristics using a BWS object case stated preference experiment designed to inform policy makers about PLWALS’ preferences for healthcare models. Attributes were developed through the expert opinion of two neurologists who specialize in ALS care (TH-P, HI), a neuroscientist who specializes in ALS research (PP), and a health economist with expertise in stated preference methods and attribute development (IH). They were all equally involved in identifying potential attributes and refining attributes within the constraints of the design considerations.

Our main design considerations were to minimize respondent burden by using as few attributes and tasks as possible. There were six objects that were features of care models. The features were length of visit (amount of time spent during a single visit), number of visits (total number of visits for ALS care within a 3-month period), number of copayments (total number of copays paid for each visit), team-based care (a group of ALS providers from different disciplines, such as neurologists, physical therapists, speech therapists, etc., working together in one location to provide your ALS care), access to clinical trials (how much knowledge the ALS care team has about clinical trials and their ability to recommend appropriate clinical trials), and ALS expertise (experience treating PLWALS).

We utilized a balanced incomplete block design (BIBD) [27]. We identified potential BIBDs that existed for a fixed number of objects (six) from a list of potential BIBDs [28] and then used the support BWS package for R to generate an experimental design for the desired criteria [29]. Respondents answered ten tasks with three objects each in which they were asked to select the most and least important items from the list (Fig. 1).

Fig. 1.

Sample best–worst scaling task

To measure quality of care, the survey included a modified version of the ALS quality measures suggested by the AAN [6]. We modified the language to make it accessible for PLWALS, as this original measure was designed for use among healthcare professionals. We asked PLWALS about their experience with each parameter. Parameters included eight universal parameters. We asked whether, within the last year, the care team developed or updated a care plan, discussed medications that slow disease progression, screened for cognitive/behavioral impairment, offered treatment for symptoms, asked about falls, and offered help with end-of-life planning. We asked whether, within the last 3 months, the care team asked about respiratory/breathing/ordered tests and whether they screened for swallowing, weight problems, or impaired nutrition. Three parameters were contingent upon clinical need. These included whether the care teams discussed options for respiratory care (contingent upon having respiratory insufficiency), offered feeding tube support (contingent upon having swallowing problems, weight loss, or impaired nutrition), and referred for communication support (contingent upon having slurred speech).

To measure access to care and satisfaction, the survey included a modified version of the Consumer Assessment of Healthcare Providers and Systems Clinician and Group Survey (CAHPS), Version 3.0, Update July 1, 2015 [30].

The survey instrument also included the Rasch Overall ALS Disability Scale (ROADS), a PLWALS-reported outcome to assess overall disability level in people with ALS. The ROADS includes 28 items, each scored as 0 (unable to perform), 1 (able to perform but with difficulty), or 2 (able to perform without difficulty) [31, 32]. The survey included the ALS-specific Quality of Life—Revised Form (ALSQOL-R), a 50-item disease-specific self-administered questionnaire in which each item is rated from 0 (least desirable) to 10 (most desirable) [33]. This includes a single-item global score, an average total score, and six domain-specific scores. The survey also captured healthcare utilization, costs, demographics, medication usage, and clinical history.

Care model types were defined and described before asking respondents to indicate which they utilized and preferred. Traditional was defined as seeing one provider per visit and requiring multiple visits to multiple doctors to receive all disciplines of care. Triage clinics were defined as specialized clinics that allow PLWALS to see multiple providers in one visit, but usually a subset of providers determined at the visit based on need. Non-triage clinics were defined as specialized clinics that allow PLWALS to see multiple providers in one visit, usually seeing all the providers in a single visit. Full definitions are provided in the ESM.

CALS Version

The CALS version mirrored the PLWALS survey to the extent possible but asked CALS to report about the PLWALS. Data were collected about the CALS as well. Caregiver Quality of Life was measured using the McGill Quality of Life in Life-Threatening Illness—Family Caregiver Version 3 (QOLLTI-F) [34]. Caregiver burden was measured using the Caregiver Burden Scale [35]. Depression and anxiety were measured using the PHQ-9 and GAD-7 Anxiety [36]. A third version of the survey was used for former CALS that eliminated questions that were not appropriate for someone who is not currently providing care.

Pilot Testing

Both surveys were pilot tested with PLWALS (n = 2) and CALS (n = 2) to check for comprehension, relevance, and appropriate respondent burden. We used a think-aloud protocol followed by a cognitive debriefing interview. Pilot respondents took the survey using a think-aloud protocol under the observation of a member of the research team (LC) [37, 38]. Using this protocol, PLWALS and CALS completed the survey in 85 and 65 minutes, respectively. Most of the changes following the pilot were minor adjustments to clarify language, improve comprehension, or offer additional answer choices for multiple choice questions to capture a fuller range of potential responses.

The survey was self-administered online via REDCap. Respondents likely completed the surveys in less time than the pilot respondents, but because they were not required to complete in one sitting, no data exist that reflects active time spent on the survey.

Data Analysis

Data were analyzed for the full sample and by respondent type (PLWALS vs CALS). This analysis combined current and former caregivers (henceforth: CALS). Comparisons were made across respondent type and care model type. Comparisons by care model type excluded respondents who were missing data about their current care model type and combined triage and traditional into one group labeled ‘all other types’.

Respondent Characteristics

Respondent characteristics include demographics and clinical characteristics of the PLWALS and are presented by respondent type. Clinical characteristics pertain to the PLWALS, regardless of the respondent.

Preferences for Care Model Types

The likelihood of choosing a specific care model was treated as a categorical variable. We report the percentage choosing each category of response (extremely likely, somewhat likely, and unlikely) for each care model type. The identification of most, middle, and least preferred care model was analyzed similarly, and we report the percentage choosing each care model type for each position (i.e., 1st, 2nd and 3rd choice). Results are presented for the pooled sample, but available by respondent type in the ESM. Chi-square tests were used to test for significant differences between PLWALS and CALS for both questions.

Prioritization of Care Model Features

The BWS data was analyzed by calculating means from a best-minus-worst analysis. A positive score indicates ‘most important’ was selected more often than ‘least important’ and a negative score indicates the opposite. We also analyzed the data using a conditional logit model, which generated the same result. Therefore, we present the results from the best-minus-worst analysis for its simplicity and easier interpretability. Results are presented by respondent type. To compare scores across respondent type, we present differences in mean scores and 95% confidence intervals resulting from t-tests to test significant differences in means.

Satisfaction and Adherence to Quality Measures

To identify potential drivers of satisfaction, we used ordinal logistic regression to estimate the relationship between overall satisfaction (dependent variable; 6-point scale from extremely dissatisfied to extremely satisfied) and quality-of-care parameters (independent variables; binary yes/no item). The model included universal parameters and excluded parameters contingent upon clinical need (e.g., excluded whether the care team discussed respiratory care because its relevance is contingent upon having respiratory insufficiency). Results for quality measures are presented for a pooled sample, as these data reflect care received by the PLWALS regardless of respondent type.

Results

The completion rate among eligible, consented potential respondents was 58.15% overall (58.12% for PLWALS and 58.22% for CALS). The analytic sample included 378 respondents, of which 254 were PLWALS (67.20%) and 124 were CALS (32.80%; composed of 79 current [20.90%] and 45 former [11.90%] caregivers).

Respondent Characteristics

The mean respondent age for the pooled sample was 60.41 years (SD 11.4); 61.09 years (SD 10.25) for PLWALS and 58.98 years (SD 12.73) for CALS (Table 1). CALS were primarily spouses or partners (n = 97, 78.23%), with some children (n = 16), other family members (n = 7), friends (n = 1), and other (n = 3). The samples were mostly white (92.86%; 93.70% for PLWALS and 91.13% for CALS) and insured by Medicare (67.32% for PLWALS and 48.39% for CALS) and/or private insurance (53.54% for PLWALS and 68.55% for CALS) and married/partnered (81.10% for PLWALS and 75.00% for CALS). The sample skewed towards higher levels of education and income; 61.03% of PLWALS and 68.55% of CALS have earned at least a college degree and about half (48.42% of PLWALS and 50.00% of CALS) have an annual income of at least US$75,000. More respondents are from the southern region of the US (i.e., Southern states) (35.83% for PLWALS; 33.87% for CALS).

Table 1.

Respondent demographic characteristics and clinical history of PLWALS

	Pooled		PLWALS		CALS
	Frequency (n = 378)	Percent (100%)	Frequency (n = 254)	Percent (100%)	Frequency (n = 124)	Percent (100%)
Age (years)
54 or less	97	25.66	60	23.62	37	29.84
55–62	90	23.81	61	24.02	29	23.39
62–68	105	27.78	80	31.50	25	20.16
68+	80	21.16	51	20.08	29	23.39
Sex
Male	162	42.86	130	51.18	32	25.81
Female	211	55.82	120	47.24	130	51.18
Race/ethnicity†
Caucasian/White	351	92.86	238	93.70	113	91.13
African American/Black	6	1.59	3	1.18	3	2.42
Asian/Pacific Islander	4	1.06	3	1.18	1	0.81
Hispanic/Latino	13	3.44	5	1.97	8	6.45
Multi-racial or other	3	0.79	1	0.39	2	1.61
Marital status
Married/living with partner	299	79.10	206	81.10	93	75.00
Divorced/separated/widowed	55	14.55	29	11.42	26	20.97
Single	18	4.76	13	5.12	5	4.03
Education, highest level completed
Some high school	3	0.79	1	0.39	2	1.61
High school/GED	46	12.17	34	13.39	12	9.68
Technical/trade	36	9.52	25	9.84	11	8.87
2-year college	46	12.17	32	12.60	14	11.29
4-year college	117	30.95	75	29.53	42	33.87
Graduate degree	123	32.54	80	31.50	43	34.68
Employment
Retired	217	57.41	160	62.99	57	45.97
Not employed	54	14.29	44	17.32	10	8.06
Employed, full-time	67	17.72	27	10.63	40	32.26
Employed, part-time	28	7.41	16	6.30	12	9.68
Homemaker	5	1.32	3	1.18	2	1.61
Student, full-time	2	0.53	0	0	2	1.61
Income ($)
≤ 25,000	21	5.56	17	6.69	4	3.23
25,000–49,999	54	14.29	35	13.78	19	15.32
50,000–74,999	61	16.14	43	16.93	18	14.52
75,000–99,999	62	16.40	38	14.96	24	19.35
100,000–149,999	59	15.61	41	16.14	18	14.52
≥ 150,000	64	16.94	44	17.32	20	16.13
Region
Southern states	133	35.19	91	35.83	42	33.87
Northeastern states	85	22.49	55	21.65	30	24.19
Midwestern states	83	21.96	61	24.02	22	17.74
Western states	61	16.14	37	14.57	24	19.35
Ex-US	12	3.17	7	2.76	5	4.03
Insurance†
Medicare	231	61.11	171	67.32	60	48.39
Medicaid	23	6.08	17	6.69	6	4.84
Other public	43	11.38	33	12.99	10	8.06
Private	221	58.47	136	53.54	85	68.55
Other	13	3.44	8	3.15	5	4.03
None/uninsured	2	0.53	9	3.54	1	0.81
Care model type
Non-triage	316	83.60	216	85.04	100	80.65
Triage	20	5.29	9	3.54	11	8.87
Traditional	28	7.41	18	7.09	10	8.06
Age of symptom onset (years)
15–49	98	25.93	71	27.95	27	21.77
50–58	96	25.40	65	25.59	31	25.00
59–65	97	25.66	68	26.77	29	23.39
66+	81	21.43	44	17.32	37	29.84
Diagnosis lag (years)
None	81	21.43	53	20.87	28	22.58
1	157	41.53	101	39.76	56	45.16
2	70	18.52	50	19.69	20	16.13
3 or more	65	17.20	45	17.72	20	16.13
Location of symptom onset†
Extremities (arms or legs)	298	78.84	193	75.98	105	84.67
Speech/swallowing	82	21.69	46	18.11	36	29.03
Breathing	21	5.56	7	2.76	14	11.29

Percentages do not sum to 100% in all categories due to missing values (i.e., skipped questions)

ALS amyotrophic lateral sclerosis, CALS caregivers of PLWALS, PLWALS people living with ALS

^†Indicates that categories for race/ethnicity, insurance, and location of symptom onset are not mutually exclusive

For the majority of the sample, (n = 316; 83.60%) the PLWALS received care via a non-triage model (Table 1). Triage and traditional models were utilized in 5.29% and 7.41% of the sample, respectively. Symptom onset began at a mean age of 56.64 years (SD 11.77) and in the extremities for most (78.84%). The average lag between symptom onset and diagnosis was 1.69 years (SD 2.52).

Preferences for Care Model Types

Respondents were asked about their likelihood of choosing each model type if given the choice. Eighty-eight percent of respondents from the pooled sample were extremely likely to choose non-triage models (87.55% of PLWALS and 89.52% of CALS; (Fig. 2). Respondents did not favor other models; 69.11% of PLWALS and 71.77% of CALS were unlikely to choose traditional models and 30.68% of PLWALS and 19.67% of CALS were unlikely to choose triage models. Differences between PLWALS and CALS (see Table S1 in the ESM) were not statistically significant for likelihood to choose non-triage (X² = 0.40; p = 0.82), triage (X² = 5.37; p = 0.07), or traditional (X² = 0.38; p = 0.83).

Fig. 2.

Likelihood to choose care model types

Respondents also rank ordered the model types in terms of preferred care model type (Fig. 3). Among the pooled sample, the first-choice model type was non-triage for 83.12% of respondents and the second-choice model type was triage for 64.63% of respondents. There were no statistically significant differences between PLWALS or CALS (see Table S2 in the ESM) for their first-choice (p = 0.14) or second-choice model (p = 0.13). The third-choice model type was traditional for both respondent groups (75.28% of the pooled respondents), but this preference was observed among fewer PLWALS compared with CALS (71.91% for PLWALS and 81.82% for CALS; p < 0.05).

Fig. 3.

Care model type rank ordered by respondent preference

Prioritization of Care Model Features

The BWS analysis for both samples indicated, relative to the other objects, the most important aspects of ALS care are ALS expertise (PLWALS score = 0.83; CALS score = 0.85) and the use of team-based care (PLWALS score = 0.45; CALS score = 0.55) (Fig. 4). Relative to other objects, the least important aspects of ALS care are length of visit (PLWALS score = − 0.35; CALS scores = − 0.38) and number of co-pays (PLWALS score = − 0.54; CALS scores = − 0.69).

Fig. 4.

Preference scores for care model features by respondent type (PLWALS vs CALS). ALS amyotrophic lateral sclerosis, CALS caregivers of PLWALS, PLWALS people living with ALS

PLWALS and CALS mostly had the same ranked order of importance for all care model features (Fig. 4). However, with regards to the number of visits, PLWALS felt this was less important than CALS (difference: − 0.12; 95% CI − 0.19 to − 0.05). This difference means that while PLWALS prioritized access to clinical trials over number of visits, for CALS these two attributes were of equal importance. There were differences in the importance of two attributes that did not impact the ranked order or differentiation among attributes. Although both PLWALS and CALS felt that the number of copays was the least important feature, compared with PLWALS, CALS felt that this feature was less important relative to other attributes (difference: 0.15, 95% CI 0.09–0.20). Both PLWALS and CALS thought team-based care was the second most important feature, but CALS prioritized team-based care relative to other features more than PLWALS (difference: − 0.10; 95% CI − 0.16 to − 0.04). Preference scores for all other objects had overlapping confidence intervals indicating no other significantly different priorities between PLWALS and CALS.

When segmented by the type of the care they currently receive (Fig. 5), team-based care was significantly more important for those in non-triage models than those for all other models (difference: − 0.21; 95% CI − 0.30 to − 0.13). Also, length of visit was significantly less important for those in non-triage than for those in all other models (difference: 0.28; 95% CI 0.20–0.38). There were no statistically significant differences in importance for the other attributes in the model.

Fig. 5.

Preference scores for care model features by care model type

Satisfaction and Adherence to Quality Measures

Overall, 89.56% (95% CI 86.40–92.72) of respondents are satisfied with their care team. Taking a more nuanced view of satisfaction, there is some indication that respondents in non-triage models may be satisfied to a greater degree than those in all other types as 53.70% (95% CI 48.12–59.27) of those in non-triage are extremely satisfied compared with 37.21% (95% CI 22.16–52.26) in all other models but the result is not statistically significant. There were no differences across model types in terms of other satisfaction measures such as ability to obtain appointments, timely responses, and follow-up on results. Overall respondents felt their care teams usually or always explained things in a way that was easy to understand (89.15%; 95% CI 86.00–92.30), listened carefully (87.04%; 95% CI 83.64–90.44), respected their preferences (88.63%; 95% CI 85.41–91.84), were knowledgeable about medical history (82.27%; 95% CI 78.41–86.14), showed respect (89.41%; 95% CI 86.30–92.53) and spent adequate time with them (85.18%; 95% CI 81.59–88.78).

Self-report of quality-of-care measures indicate some significant differences between groups (Fig. 6). Non-triage PLWALS are more likely to have an MDC care plan developed or updated than PLWALS in all other model types (mean difference: 31.26; 95% CI 17.48–45.05), and more likely to be asked about falls (16.04; 95% CI 8.46–23.62) and offered help with end-of-life planning (21.96; 95% CI 6.21–36.41). Further, non-triage PLWALS are more likely to be queried about breathing or have tests ordered (18.68; 95% CI 7.33–30.03), referred to speech specialists (22.00; 95% CI 7.21–36.70), and screened for swallowing/weight loss/nutrition (24.38; 95% CI 11.86–36.91). They are also more likely to be screened for cognitive and behavioral impairments (16.31; 95% CI 1.13–31.49) and offered treatment for symptoms (18.18; 95% CI 3.22–33.13), although these two parameters were only marginally significant.

Fig. 6.

Adherence to AAN quality measures (as self-reported by respondents). AAN American Academy of Neurology. *Indicates statistically significant difference in means based on confidence intervals. ^†Indicates that parameter was assessed conditional on clinical appropriateness (e.g., asked about whether offered traditional for speech therapist only if they had been informed of slurred speech)

While there was no difference in overall satisfaction between care model groups, the analysis to identify associations between quality-of-care measures and overall satisfaction resulted in two significant items, holding other quality-of-care items constant. The odds of being extremely satisfied were 2.31 (95% CI 1.47–3.64) times greater for those whose care team developed or updated a care plan within the last year compared with those whose care team did not. The odds of being extremely satisfied were 3.16 (95% CI 1.38–7.27) times greater for those who were asked about falls than for those who were not asked. All the other quality-of-care items were not statistically significant.

Discussion

In recent years, emphasis has been placed on patient-centered healthcare, including patient satisfaction and patient preferences for healthcare. This study is directed at identifying what PLWALS and CALS consider the most important aspects of their care, specifically how they value the features of MDC. Previous preference studies that examine preferences in healthcare delivery models have estimated the importance of MDC relative to other features of health care delivery [39] or have excluded MDC from the attribute list [40]. The current study is novel in its specific focus on the relative importance of MDC features as well as the application to ALS.

Respondents of this survey were largely utilizing a MDC model, specifically a non-triage MDC model. Thus, the results indicating a preference for non-triage MDC and the value placed on its features should be interpreted with caution, as this is a potential source of bias. Nonetheless, respondents clearly preferred a non-triage model if given the choice. This model includes multiple specialists organized as a team(s) with ample experience and expertise in managing the disease and its symptoms, all seeing the PLWALS in one visit. This preference is consistent with the result of the most important features of their care being ALS expertise and team-based care. Comprehensive and personalized care for PLWALS requires coordinated knowledge and attention by multiple specialties. This finding is also consistent with other studies that found cancer patients value seeing multiple providers in one visit [41] and value their specific care team members [42].

Although not surprising that the aspects of care related to the providers (ALS expertise and team-based care) were top priorities, it is interesting that clinical trial access was not a top priority given the few treatments available for ALS. The process-oriented attributes of the care models, like visit frequency, visit length, and cost were less important. Although this is inconsistent with previous findings that cost is a most important attribute generally, our findings are consistent with prior work that has shown that the importance of cost is inversely related to the severity of disease [39]. Our study indicates that PLWALS may be willing to pay more, make more frequent visits or have longer visits to assure clinical expertise and team-based care. This has direct implications for clinical practice in terms of team composition and clinic administration to facilitate collaboration.

The financial aspects of having multiple co-pays were least important from the perspective of PLWALS in the current study. From the providers’ perspective, the reality of managing MDC clinics requires flexible and creative approaches in keeping ALS clinics financially solvent in the current medical model [14], which does not distinguish between a single provider and a team-based approach. The current model also does not reward institutions who offer the non-triage model over the other two options. Except for the Veteran’s Affairs (VA) system, which is organized differently, the maintenance and delivery of the non-triage model is largely left to the resourcefulness and efforts of ALS providers who frequently need to rely on philanthropy [14]. The extent of this problem is the topic of an ongoing study.

Further, PLWALS and CALS utilizing non-triage MDC models observed more adherence to quality care measures. The non-triage model outperformed the other two models in areas such as regular respiratory testing, swallowing and nutritional care, feeding tube discussion and support, end-of-life planning, cognitive and behavioral assessment, etc. This supports the importance of providing adequate access for PLWALS and CALS to non-triage MDC care in underserved areas and providing resources to MDC clinics to make it feasible to provide non-triage type care [14]. However, improving access to non-triage MDC care may result in greater guideline-concordant care, including additional visits or testing. An unintended consequences of higher healthcare utilization may be greater financial toxicity, as has been observed among people with lung cancer [42].

Other studies have found that patients with chronic conditions in MDC have reported greater satisfaction with care than those not receiving MDC [43–45]. We found no evidence of significant differences in satisfaction across care models, but that may reflect a lack of awareness about the possibility for other care models that PLWALS have experienced. Importantly, people may be satisfied with the status quo, but when asked would they choose non-triage MDC if given the option, most answered yes. We identified two quality-of-care items that are significantly associated with higher satisfaction: use of a care plan and asking about falls. Both items were addressed more sufficiently in the non-triage model. This indicates that there are potential gains in both satisfaction and quality of care with more comprehensive MDC models where PLWALS are assessed by all disciplines at every visit. More research is needed to assess care preferences more widely and to understand the relationship between types of care and quality-of-life outcomes.

Differences in adherence to quality-of-care measures, regardless of statistical significance, are clinically meaningful. Differences in adherence to quality-of-care measures represent the possibility that important aspects of care may be overlooked. This is a key differentiator between non-triage and triage models. In a non-triage model, because the PLWALS see every specialist every visit, they are questioned about all the discipline-specific issues that are relative for their current disease stage. However, in a triage model, the care delivered is determined by what the screening medical professional perceives the needs of PLWALS are on a given day, but subject to variability based on the effectiveness of the triage assessment.

Limitations

Our findings are likely influenced by the limitations of the study design. First, ALS diagnosis (or presumed diagnosis) was self-reported and cannot be confirmed. Second, the survey was distributed widely via email but there is inherent selection bias which may limit the generalizability of the findings. Respondents were recruited through organizations that maintain distribution lists that may be biased towards those PLWALS and CALS who are likely more activated and motivated to obtain disease-related information. Certainly, results are biased towards those earlier in their disease progress and without significant cognitive impairment. The average ROADS functioning score was 71.95 (SD 23.14) out of a maximum score of 146 [31]. In addition, our respondents do not match the demographic distribution of PLWALS in the US. Further, this survey was biased towards those able to access the internet and underserved areas are likely under-represented. Our sample was largely from those currently using non-triage models, which could bias the preference towards that model. Nevertheless, the fact that both PLWALS and CALS preferred the non-triage model clearly underlines the benefits that this type of care provides.

A third limitation of this study is that we measured adherence to AAN practice parameters using quality-of-care measures that are not intended to be patient-reported. Although not validated, we intended to capture the perspective of PLWALS about items that the AAN has deemed to be important for ALS quality care. For some of these parameters, PLWALS may not be aware of how the care they receive aligns with or accomplishes the parameter. For instance, if parameters are discussed using different language or the purpose of the parameter is not made explicit to PLWALS and CALS, they might not recognize the care as satisfying the parameter. Furthermore, these measures relied on the respondents’ recall. Nonetheless, any discrepancies resulting from recall bias would be equally likely to occur across model care types and therefore would not bias the differences between groups.

Finally, attributes for the BWS experiment were developed without conducting formative qualitative work and relied heavily on expert opinion. While this may limit the validity of the evidence, our goals were narrowly focused on a limited set of features that varied by care model type and thus didn’t necessitate broad exploration of new themes.

Conclusion

PLWALS and CALS prefer non-triage models of MDC and prioritize attributes of non-triage models, such as ALS expertise and team-based care. Survey respondents using a non-triage model also perceive greater adherence to important quality-of-care indicators than those using other models. These findings should be interpreted with caution, as most study respondents were receiving care via a non-triage model. Nonetheless, these findings underscore the importance of MDC in providing the best quality care that is patient-centric. This knowledge can help clinicians align care with patient values. It can also empower decision makers to support MDC models as it highlights the need to provide resources that will support this approach to ALS care. AAN can use these findings to provide more detailed policy recommendations about care delivery models. The greatest impact may be realized by payers incentivizing centers to adopt non-triage models through policies that increase reimbursement rates for MDC providers or the adoption of pay-for-performance models that reward adherence to quality metrics. These policy recommendations have the potential to contribute to the financial viability of non-triage care models, which will become increasingly important as ALS prevalence rises in the near future.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Funding

This study was funded by Mitsubishi Tanabe Pharma America as part of a grant titled “Characterizing outcomes, costs and preferences for treatment settings for patients with ALS and their caregivers.”

Data availability

Anonymized data that supports these findings may be made available by the authors upon reasonable request from any qualified investigator.

Author contributions

Terry Heiman-Patterson and Ilene Hollin conceptualized the overall study. All authors contributed to the design of the instruments and contributed equally to the acquisition of data. Ilene Hollin conducted the data analysis and drafted the original draft manuscript. All authors commented on previous versions of the manuscript and all authors read and approved the final manuscript.

Competing interests

Ilene Hollin is a member of the editorial board of The Patient. She was not involved in the selection of peer reviewers for the manuscript nor any of the subsequent editorial decisions. She received salary support from Mitsubishi Tanabe Pharma America for this study. She receives research support that is unrelated from the NIH (K01HL161330). Hristelina Ilieva is a site-PI for the Healey trial, receives research support from Mitsubishi Tanabe Pharma America, Sanofi, and Farber Family Foundation. Piera Pasinelli is supported by grants from NIH (RO1 NS136624; R21 AG093032); ALSUM; Aldrich Foundation; Family Strong4ALS, and Farber Family Foundation. L. Chisholm is supported by funding from Mitsubishi Tanabe Pharma for this grant and has no other relevant disclosures. T. Heiman-Patterson is supported by Mitsubishi Tanabe Pharma America. She receives unrelated funding as site PI for clinical trials from COYA therapeutics, Amylyx, Arcellx, and the Healey Platform trial. She is site PI for All ALS and for the Natural History Consortium.

Ethics approval

This study was approved by the Institution Review Board of Temple University (protocol #29392). The IRB determined that written consent was not required as responding to the survey implied participant consent.