Exploring Treatment Expectations and Clinical Meaningfulness in Spinal Muscular Atrophy Using the Goal Attainment Scale

Abstract

Objectives

To investigate how Goal Attainment Scale (GAS) in spinal muscular atrophy (SMA) identifies patient-reported clinically meaningful improvement after disease-modifying treatment and to evaluate whether the patient-stated goals identified through GAS are already captured by items in existing motor assessment scales.

Design

Longitudinal retrospective study considering the first available follow-up.

Setting

The study was conducted in a rehabilitation center dedicated to neuromuscular diseases.

Participants

The GAS was recorded from N=58 SMA patients on treatment with nusinersen or risdiplam regularly followed for clinical follow-up in 2023.

Main Outcome Measures

Clinical and demographic data, SMA type, Hammersmith Functional Motor Scale, Children’s Hospital of Philadelphia scale, Revised Upper Limb Module, 6 Minute Walk Test, and GAS questionnaire were collected.

Results

One hundred forty-nine SMART goals were clustered in 15 macrodomains, with the most frequent being mobility, upper limb strength, and activities of daily living. Most smart goals (72%) could be redirected to items of standardized motor and functional scales. Unmet goals primarily involved activities of daily living, endurance, and mobility, with a particular focus on upper limb strength.

Conclusions

The GAS provides an individualized, reliable, and quantifiable measure of goal attainment, and is useful in setting realistic treatment expectations that could guide patients and physicians to interpret changes over time.

Highlights

• •The Goal Attainment Scale showed good reliability, helping tailor individualized treatment goals.
• •Fifty-eight spinal muscular atrophy patients set 149 Specific, Measurable, Attainable, Realistic, and Timely goals, mainly about movement, strength, and activities of daily living.
• •72.48% of patient goals aligned with existing neuromotor functional scales.
• •28% of patient goals were not captured by the existing assessment tools.

Spinal muscular atrophy (SMA) is an autosomal recessive neurodegenerative disorder caused by mutations in the SMN1 gene, encoding for the survival motor neuron protein, which leads to progressive degeneration of alpha motor neurons and consequently to generalized hypotonia, muscle atrophy and weakness.¹

There are now 3 approved disease-modifying treatments for SMA, and whereas the pivotal studies focused on pediatric onset^2,3 and on the most severe forms of SMA, there is now increasing evidence from randomized controlled trials^2,3,4 and real-world data^5,6,7 that treatment with SMN2-dependent drugs also provides benefits to adults or at least stabilizes and reduces their rate of decline. Standarized functional scales usually quantify disability and residual abilities based on predefined tasks and achievements. However, adult patients living with SMA are a very heterogeneous population⁸ and preset goals using the standardized functional scales may not be able to detect changes in all patients and may not capture what matters most to these patients. Skills such as finger movement, bulbar functions, endurance, independence, and muscle strength changes are only partially detectable by these scales.

For this reason, patient-reported outcomes may represent a valuable complementary approach, because they provide the patients perception of improvement in specific tasks of everyday activities which are relevant to patients, and are in fact used in expert centers together with the neuromotor and other assessment scales.⁹

A large survey conducted by Mongiovi et al¹⁰ did collect the most relevant themes for patients living with SMA and did rate how much each of these impacted the patients’ perception of burden. Based on this study, the SMA-Health Index (SMA-HI) was in fact developed to capture a measure of disease burden as perceived by patients and to see how this may change over time with treatment or as disease progresses.^10,11 Yet the SMA-HI does not provide, because it was not the intent of the scale, data on patients’ expectations and specific goals which are often needed to tailor and manage expectations.

The Goal Attainment Scale (GAS) was first introduced in 1960s by Kiresuk and Sherman¹² as a person-centered approach¹³ assessing individual goals and rehabilitation outcomes^14,15. GAS goals should be, as much as possible, Specific, Measurable, Attainable, Realistic and Timely (SMART).¹⁶ The GAS has been shown to be reliable, responsive, and sensitive to changes,17, 18, 19 and supports patient-centered care²⁰. Interest in the GAS is also growing in the SMA community, but the experience is still limited.²¹

In this study we wanted to investigate how the application of the GAS in SMA helps to identify patient expectations regarding clinically meaningful improvement after disease-modifying treatment at baseline and over time. In addition we also wanted to evaluate whether the patient-stated goals identified through GAS are already captured by items in existing motor and functional assessment scales.

Methods

Study population

Charts from SMA patients on treatment with nusinersen or risdiplam referred to our Clinical Centre (Milan) between December 2018 and May 2023 were retrospectively reviewed. All patients signed an informed consent to participate (MER-SMA-18-003).

For each patient, clinical and demographic data, SMA type, motor, nutritional, and respiratory status, and SMA-HI and GAS questionnaires were collected at baseline and at the first available follow-up visit. A subgroup of patients on stable treatment were asked to repeat the GAS interviews after 10-14 days from their first one to test for reliability.

The Goal Attainment Scale

The GAS was administered by a dedicated psychologist with experience in SMA at baseline and at follow-up during out-patient visits. As per GAS instructions, each patient had to provide 3 SMART goals in any of several domains which were given as an example (motor, respiratory, nutritional, psychological, etc) and which were discussed with the psychologist and referring adults and pediatric neurologist at baseline and at follow-up visits, once provided by the patients during their visit. A baseline score of either −2 or −1 was provided by each patient to describe the perceived difficulty in reaching that specific goal; whereas the achievement of the goal was rating on a 5-point scale (−2, −1, 0 +1, +2) in which the scores −2 and −1 were given to indicate that the patient achieve much less (−2) or somewhat less (−1) than the expected outcome, a score of 0 was given to indicate that the patient perceived full achievement in that goal, and the scores of +1 and +2 were given if the patient achieve somewhat more (+1) or much more (+2) than the expected outcome.

Clinical assessments

Patients were subjected to standardized neuromotor assessments as appropriate for age and level of disability. In specific, Hammersmith Functional Motor Scale (HFMSE, age range: ≥2y, applicable to sitters and walkers); Revised Upper Limb Module for upper limb function (RULM, age range: ≥2.5y); Children’s Hospital of Philadelphia (CHOP INTEND SCALE, age range: no age restriction) and the 6 minute Walk Test (6MWT, age range: no age indicated). Any side-effects from treatment were also recorded.

SMART goals and corresponding items in the standardized motor and functional scales

To verify whether the SMART goals identified by patients were already included in the neuromotor function scales used routinely to monitor disease progression in SMA, the physiotherapist and the referring physicians verified whether each SMART goal could be attributed to a specific item or task in one of the exitisting motor and functional scales as shown in supplemental table S1 (available online only at http://www.archives-pmr.org/). If there was a corresponding item in the neuromotor functional scales, that specific SMART goal was rated as “covered” by the existing scales; if not, the SMART goal was rated as “missing.”

Statistical analysis

The reproducibility of the scale was evaluated considering a test-retest analysis based on patients’ concordance in defining goals in a 2 weeks range of time (same goals reported by the patient at 2wks interval). Concordace was defined as “full” (the patient chose the same 3 goals), partial (the patient repeats some goals but changes others) or absent (the patients chose 3 different goals), and was expressed in terms of numbers and percentage. It should be noted that for those cases where there was a mismatch between patient perception and objective measures, the goals defined during the first assessment were used for subsequent analyses. Changes in the clinical motor scales were classified as worsened/stable/improved in agreement with the minimum clinically important difference established for each scale and adopted in clinical trials. Minimum clinically important difference thresholds reported in the literature were applied at the individual level; therefore, given the discrete nature of the scales, cut-offs originally expressed as decimal values were operationalized using the nearest integer values fulfilling the original inequality criteria. In detail, a change of at least ±3 points in HFMSE for SMA1, a worsening of ≤−4 points or an improvement of ≥+2 points for SMA2, a worsening of ≤−4 points or an improvement of ≥+3 points for SMA3, a change of at least ±2 points in RULM, at least ±4 points in CHOP, and at least ±30 meters in the 6MWT were considered as cut-offs to classify changes.22, 23, 24, 25, 26, 27, 28

In order to qualitatively assess the concordance between the changes detected by the clinical motor scale and those detected by the GAS scale, the SMART goals outcome scores were not incorporated into the single aggregated T score,¹⁴ but were considered as raw scores, assessing scores of −2 and −1 as worsened, 0 as stable, and +1 and +2 as improved. When the patient goal could be related and transferred to a score on at least 1 item of the objective functional measures a match (concordance) or mismatch was recorded.

The concordance was defined as full concordance when both the motor scale and the GAS classified the changes equally (both as worsened, both as stable, or both as improved), as partial concordance when 1 of the 2 measurement methods (clinical assessment or GAS) classified the change as stable, whereas the other one classified it as worsened or improved, and as mismatch when 1 of the 2 measurement methods classified the change as worsened, while the other one classified it as improved, or vice-versa. The concordance in each comparison was analyzed through contingency tables and expressed in terms of frequencies and percentages. The sampled population was tested for concordance without any stratification, eg, in pediatric and adult cohorts, in line with the objective of the study which was not related to an overview of the actual patients’ needs but based on the ability of GAS in detecting motor changes based on patient’s perception regardless of who actually expressed the need (eg, patient or patient’s parents).

Statistical analyses were performed using SAS 9.3 (SAS Institute Inc).^a

Results

A convenience sample size of 58 SMA patients were included in the study, with a median age at first evaluation of 18.48 years [8.12-32.88], and a men/women ratio of 1.23 (32 men / 26 women). Table 1 summarizes the descriptive characteristics of the SMA cohort. The average time between the first GAS administration (T0) and the first follow-up (T1) was 10.62 months [5.60-21.08].

Table 1.

Demographic and clinical characteristics of the enrolled SMA patients at first GAS administration (n=58).

Age, y	18.48 (8.12-32.88)
Sex, n (%)
Male	32 (55.17)
Female	26 (44.83)
SMA type, n (%)
I	12 (20.69)
II	29 (50.00)
III	17 (29.31)
Motor function, n (%)
Nonsitter	27 (46.55)
Sitter	25 (43.11)
Walker	6 (10.34)
Treatment, n (%)
Risdiplam	17 (29.31)
Start at baseline, n (%)	11 (64.71)
<12 mo, n (%)	6 (35.29)
Spinraza	41 (70.69)
Start at baseline, n (%)	5 (12.20)
<12 mo, n (%)	19 (46.34)
<24 mo, n (%)	17 (41.46)
Ventilation status, n (%)
No support needed	23 (50.00)
NIV≤8 h/d	6 (13.04)
NIV>8 h/d	10 (21.74)
NIV 24 h/d	1 (2.17)
IV	6 (13.04)
Missing	12
Nutritional status, n (%)
Gastrostomy	9 (15.52)
Modified diet	9 (15.52)
Free diet	40 (68.97)
Scoliosis, n (%)
Yes	39 (78.00)
Requiring surgery	13 (33.33)‖
No	11 (22.00)
Missing	8
Motor scale
HFMSE, n=46*	4.00 (0.00-26.00)
RULM, n=41†	15.00 (4.00-31.00)
CHOP INTEND, n=12‡	18.50 (7.00-47.00)
6MWT, n=5§	200.00 (100.00-530.00)

NOTE. Data are expresses in terms of median [interquartile range], except where otherwise indicated.

Percentages were calculated based on available data; missing data are reported but not included in percentage calculations.

Abbreviations: 6MWT, Six Minute Walking Test; CHOP INTEND, Children’s Hospital of Philadelphia Infant Test of Neuromuscular Disorders; HMFSE, Hammersmith Functional Motor Scale Examination; IV, invasive ventilation; NIV, noninvasive ventilation; RULM, Revised Upper Limb Module.

^⁎HFMSE scores ranges from 0 to 66. The higher scores indicate better motor function.

^†RULM scores ranges from 0 to 37. The higher scores indicate better motor function.

^‡CHOP scores ranges from 0 to 20. The higher scores indicate better motor function.

^§6MWT: more than 300 mt indicated a normal function.

^‖All the required interventions were spinal surgery.

Considering a maximum of 3 goals per patient, 149 unique SMART goals were established. Specifically, 33 patients set 3 goals, whereas 25 set 2. Each goal was weighted and consequentially categorized by 3 NeMO clinicians with expertise in neuromuscular diseases, including SMA, who identified 15 goal domains based on the characteristics of each SMART goal. Table 2 summarizes frequency and percentage of the SMART goals grouped by domains on both the overall sample and stratified by SMA function. In the overall population, the most frequently reported goals were related to independence in movement, upper limb strength, and activities of daily living: a pattern also observed among sitter patients. Among nonsitter patients, the most common goals concerned upper limb strength, activities of daily living, and breathing, whereas walker patients most frequently reported goals targeting independence in movement, endurance, and upper limb strength (table 2).

Table 2.

Frequency and percentage (in brackets) of the SMART goals grouped by macrodomains.

		SMA Function			Baseline Rating†
MACRODOMAINS	Overall	Nonsitters	Sitters	Walkers	−2	−1
Independence in movement	40	3 (7.50)	29 (72.50)	8 (20.00)	13 (32.50)	27 (67.50)
Upper limbs strength	37	22 (59.46)	13 (35.14)	2 (5.41)	9 (24.32)	28 (75.68)
Activities of daily living	22	9 (40.91)	13 (59.09)	0 (0.00)	5 (22.73)	17 (77.27)
Endurance	13	8 (61.54)	2 (15.38)	3 (23.08)	0 (0.00)	13 (100.00)
Breathing*	12	9 (75.00)	3 (25.00)	0 (0.00)	2 (16.67)	10 (83.33)
Neck strength	7	5 (71.43)	2 (24.57)	0 (0.00)	0 (0.00)	7 (100.00)
Wheelchair control	6	5 (83.33)	1 (16.67)	0 (0.00)	1 (16.67)	5 (83.33)
Swallowing and nutrition	4	4 (100.00)	0 (0.00)	0 (0.00)	0 (0.00)	4 (100.00)
Verbal communication skills	2	2 (100.00)	0 (0.00)	0 (0.00)	0 (0.00)	2 (100.00)
Facial expression	1	1 (100.00)	0 (0.00)	0 (0.00)	0 (0.00)	1 (100.00)
Fine motor skills	1	0 (0.00)	1 (100.00)	0 (0.00)	0 (0.00)	1 (100.00)
Scoliosis	1	0 (0.00)	1 (100.00)	0 (0.00)	0 (0.00)	1 (100.00)
Lower limbs strength	1	0 (0.00)	0 (0.00)	1 (100.00)	0 (0.00)	1 (100.00)
Pain	1	1 (100.00)	0 (0.00)	0 (0.00)	0 (0.00)	1 (100.00)
Plague management	1	1 (100.00)	0 (0.00)	0 (0.00)	0 (0.00)	1 (100.00)

^⁎Breathing: use of the ventilator for less hours per day.

^†Frequency and percentage (in brackets) of patients who rated the difficulty in reaching their goal as moderate (−2) or slight (−1).

Abbreviations: SMA, spinal muscular atrophy; SMART, Specific, Measurable, Attainable, Realistic, and Timely.

Considering independence in movement, upper limb strength, and activities of daily living (ADLs) as the 2 macrodomains areas most frequently cited across the 3 clinical groups classified by residual motor function (walkers, sitters, and nonsitters), it emerges that within each macrodomain, the specific goals differ according to the level of residual functionality.

With regard to independence in movement, walkers primarily identify goals related to stability and locomotor tasks (eg, “Maintain stability while standing” and “Be able to climb stairs”), sitters emphasize goals concerning postural transitions and rotational movements (eg, “Postural transition between bathroom/wheelchair/bed” and “Be able to turn by oneself”), while nonsitters focus on achieving an autonomous seated position (eg, “Sit independently”).

In the domain of upper limb strength, walkers prioritize goals aimed at improving grip strength and upper limb function to enable independent standing from the wheelchair (eg, “Regain grip in the left hand” and “Have enough strength in the arms to stand up from the wheelchair independently”). Sitters tend to focus on enhancing muscle strength for object lifting (eg, “Lift the glass to drink” and “Be able to lift 1 kg with the right hand”), whereas nonsitters primarily target distal functions and the ability to manipulate small objects (eg, “Grab small toys on the right and left when sitting” and “Press the keyboard keys”).

Finally, in the ADL macrodomain, whereas sitters report a wide range of goals (eg, “1h a wk to be able to go to the mall with friends” and “Dress by myself”), nonsitters predominantly emphasize goals related to autonomous eating (eg, “Eat independently”).

The reproducibility of the scale in terms of patients’ stability was evaluated in a subgroup of 23 patients who were subjected to the scale a second time after a 2-weeks’ period. Table 3 summarize the concordance between the 2 evaluations.

Table 3.

Qualitative analysis of the reproducibility of the GAS.

	No. (%)	Interpretation
Full concordance	18 (78.26)	SMART goals that were categorized in the same domain as the goals selected 2 wks earlier
Partial concordance	2 (8.70)	SMART goals that were only partially classified in the same domain as the goals selected 2 wks earlier
No concordance	3 (13.04)	SMART goals that were categorized in different domain as the goals selected 2 wks earlier

The qualitative analysis of concordance between improvement, stability or worsening in GAS pre-selected goal domains, and the corresponding clinical improvement, stability, or worsening in the related motor scales (HFMSE, RULM, CHOP, and 6MWT) is shown in figure 1. In addition, figure 1 also includes the list of the specific goal domains chosen in accordance with the related motor scales to which they have been associated.

Fig 1.

Concordance (full and partial) and discordance between GAS and standardized functional scales.

Of the 149 goals reported by individuals with SMA, 108 (72.48%) were “covered” by the existing validated motor and functional scales for SMA, because the goals set by the patients could find a correspondence in at least 1 item of the motor and functional scales in accordance with the age of patients, whereas 41 (27.52%) were “missing.” These “missing” goals were more related to ADL, endurance, and independence in movement with a particular focus on upper limb strength. Supplemental table S1 lists the single SMART goals given by each patient and shows the attribution to a specific task or measure in the motor and functional scales used routinely in clinical practice in SMA, whereas table 4 summarized the “missing” goals only.

Table 4.

Frequency and percentage of the missing uncovered SMART goals.

Fields	No. (%)	Example
Independence in movement	11 (26.83)	"Improve movements in water"
Endurance	10 (24.39)	"Increase endurance during hockey games"
ADL	9 (24.39)	“Take off t-shirt independently”
Upper limb strength	6 (12.20)	“Stronger grip with the hand”
Neck strength	2 (4.88)	“Turn the neck better to the left when sitting on wheelchair”
Pain	1 (2.44)	“Lift the head without pain”
Plague management	1 (2.44)	“Continue to remain free of bedsores”
Swallowing and nutrition	1 (2.44)	“Swallow meat without needing to add sauce”

Discussion

Our data support the use of the Goal Attainment Scale in patients with SMA as a tool to capture their expectations and quantitatively assess the extent to which individual goals are met.

When the SMART goals were clustered in 15 macrodomains by NeMO clinicians, not unexpectedly, the most frequent goals were those pertaining to independence in movement (40 goals), upper limb strength (37 goals) and ADL (22 goals). In general, there was a good correspondence between goals set by the patients and their disability level and residual function. Sitters more frequently chose goals aimed at improving wheelchair mobility or achieving some degree of independence in “sit-to-stand” movements. Nonsitters, on the other hand, mainly selected goals related to the maintenance or improvement of respiratory function or nutritional status, as well as goals concerning wheelchair use and control (eg, finger movements for joystick use, upper limb strength, or neck control) or specific domains such as communication or facial mimic. Walkers, consistently with their residual abilities, chose goals focused on improving endurance or preserving ambulatory status. Overall, these findings suggest that, within a structured goal-setting process supported by an experienced neuropsychologist, patients tended to identify goals focused on improving existing skills rather than acquiring entirely new abilities. Professional guidance played a key role in ensuring that goals were realistic, coherent, and consistent with patients’ level of disability. These results show that patients’ functional goals are closely shaped by their residual motor abilities, emphasizing the importance of individualized goal-setting and tailored interventions to ensure that rehabilitation strategies align with both functional status and daily priorities.

Most items in the goals set by patients were covered by items in the motor and functional scales creating the framework for the physicians to better focus on the potential benefits of treatment on an individualized basis. The goals set by the patients could be realistically checked when applying the functional scales in routine practice, and could therefore clearly measure expectations and responses, and strengthen the clinical significance of any given treatment for patients. Moreover, when comparing patient perception of change with respect to a specific goal and the objective findings on the motor and functional scales, there was full or partial concordance between the subjective perception the patients had and the objective changes documented in the motor and functional scales. This supports the fact that the expectations emerging from the GAS and discussed with the referring physician may actually be verified and documented by the objective neuromotor scales. This may potentially facilitate the discussion between patients and prescribing physicians when discussing treatment with each single patient because individualized goals may be set and tailored to be accomplished in a more specific way. However, there were a minority of cases where the patient perceived the goal was not reached because there was a perception of worsening on a specific function but in fact the objective assessments showed no change; or, in other cases there was a perception of improvement whereas the objective measures showed no change. These cases provide additional evidence that the patients perception may not match the objective results of the functional scales but this does not necessarily mean that the patient perception is false and irrelevant. It may be the patient has either gained or lost compensatory mechanisms, has changed posture or has adapted to specific situations without modifying the scores on the motor or functional scales. On the other hand, there were cases in which the patient perceived there was no change in the goals set yet the motor or functional scales did change. This may reflect the fact that patients may not perceive numerical changes as clinically meaningful to them, once again strengthening the idea that subjective and objective measures should be considered as supportive elements to interpret treatment benefits. Of note, there were no cases in whom the patients reported to have worsened while the neuromotor scales showed improvement in their scores. Interestingly there were 6 cases in whom there was a complete mismatch in the opposite direction, these being cases where patients reported an improvement in the GAS while the motor or functional scales actually reported worsening of the scores. For example, a patient who worsened in the elbow flexor muscles so that his scores on the RULM worsened by 2 points the mismatch could be the result of the fact that the patient can compensate for that loss, so that there is no perception of actual worsening. Another possible explanation for the mismatch was referred to a patient who did not change or worsen in the neuromotor scales but felt that he perceived improvement in T other domains, like speech or swallowing for instance which are not captured by the neuromotor scales.

There were a minority of “missing” goals which could not be attributed to any of the motor or functional scales or to any other known outcome. For example, specific functions involving the upper limbs were rated as clinically meaningful and were classified as “missing” in the existing motor and functional scales. These referred to functions involving improved finger dexterity or proximal muscle strength which were initially not even contemplated as a possible target of improvement by patients but which have become common expectations among patients since the advent of disease-modifying treatments.

Study limitations

Although promising and interesting, this study is exploratory. One of the advantages of the GAS is that it is individualized and person-centered. Yet at the same time it does not allow standardization of results to best interpret the findings. The clusterization of the goals is one way we have tried to overcome this limitation. In this pilot study, the referring physicians, the physiotherapist and the psychologist all worked together to support and assist patients in defining their individual objectives and set their expectations, which need to be realistic and targetable. In clinical practice, we expect that, with appropriate training and a standardized operating procedure with specific reference to SMA, the GAS will be administered in a more standardized and structured way. A working hypothesis is currently exploring the possibility of establishing 'item banks’ of goals as suggested by Tennant,²⁹ in order to reduce the possibility that the patient chooses non-SMART goals.

Conclusions

The GAS provides an individualized, reliable, and quantifiable measure of goal attainment, supporting the setting of realistic treatment expectations. In our study, 72% of SMART goals were captured by existing assessment tools, whereas 28% were not, suggesting there may be a need to revisit or implementcurrent measures. These findings support the use of the GAS as a complementary tool alongside motor and functional scales and other tests routinely used in clinical practice. The GAS may support physicians to define responders and less-responders. So far, a less-responder is defined as such only based on the clinically meaningful significance on the neuromotor scores,22, 23, 24, 25, 26, 27, 28 which although crucial to define response to treatment, may not correspond to the perceived benefit a patient may have in other functional domains or in some daily activities which may not be accounted for by the motor and functional scales. The application of the GAS in routine care in larger cohorts of patients with SMA and over time will verify whether it may actually be a useful tool to set patient goals and expectations and facilitate patient-physician alliance in defining treatment strategies and outcomes.

Supplier

a. SAS, version 9.3; SAS Institute.

Disclosure

J.C., G.C., V.A.S., and E.A. reports personal fees from BIOGEN S.R.L., ROCHE, NOVARTIS outside the submitted work; Mercuri E. reports personal fees from BIOGEN S.R.L., ROCHE, AVEXIS, SCHOLAR ROCK and NOVARTIS outside the submitted work. The other authors have nothing to disclose.