Enfeeblement in Essential Tremor: Development and Validation of a New Psychometric Measure

Abstract

Introduction:

Despite a growing body of knowledge regarding essential tremor (ET), past studies have fallen short in capturing the full impact of ET on patients and caregivers. We propose enfeeblement (i.e., having the qualities of being prematurely old, helpless, or debilitated) as a novel clinical outcome measure in ET. Due to the lack of enfeeblement scales for ET in the literature, we developed and validated an enfeeblement scale for ET.

Methods:

The Essential Tremor Enfeeblement Survey (ETES) consists of eight 5-point Likert-type scale questions and is designed to be a caregiver-reported outcome.

Results:

Enfeeblement scores showed a floor effect of 15.3%, no ceiling effects, and demonstrated good overall test–retest reliability (intraclass correlation coefficient = 0.73), favorable internal consistency (Cronbach α coefficient = 0.92) and good convergent validity.

Conclusion:

The ETES has robust properties. Aside from future studies of enfeeblement in ET, enfeeblement should be explored more broadly as a psychometric measure across other neurological disorders.

Introduction

Essential tremor (ET) is among the most prevalent neurological disorders and the most common form of abnormal tremor in humans. Aside from kinetic tremor, patients with ET may manifest additional motor and nonmotor features.^1,2 Motor features include a range of tremors (e.g., postural, kinetic, intention, resting) as well as mild gait ataxia.^3-5 Nonmotor features may include cognitive impairment (from mild cognitive difficulty to dementia),^6,7 anxiety,⁸ depressive symptoms,^9,10 and sleep dysregulation.^10-12 Tremor in ET may result in social avoidance¹³ and embarrassment,¹⁴ reducing quality of life.¹⁵ The tremor of ET may also be severe enough to impact daily activities, producing functional disability.^16-18

Typically, impairment due to ET is not severe enough to require constant care or paid caregivers; however, it may require the assistance of relatives or friends who serve as informal caregivers.¹⁹ Such care may be associated with some degree of caregiver burden.

We recently examined the prevalence and correlates of caregiver burden in ET,^19,20 finding that 11% of caregivers provided >25 hours of care each week, and 13% experienced high levels of caregiver burden.¹⁹ Although we were able to identify several important sources of caregiver burden (e.g., cognitive problems, falls, and depressive symptoms among ET patients),¹⁹ it was clear that our models fell short of capturing all sources of caregiver burden. As a result, we conjectured the presence of an additional source(s) of caregiver burden. During subsequent discussions with caregivers of ET patients, a theme began to emerge that caregivers with higher amounts of caregiver burden viewed their loved ones as having the qualities of being prematurely old, slightly helpless, or debilitated. We began to conceptualize this overall quality as one of “enfeeblement,” for which there are no scales to measure it in the literature. Here, we describe how we developed and validated a survey designed to measure enfeeblement in ET. Although the scale was developed for ET, we recognize that it could be modified to assess enfeeblement more broadly across a range of debilitating neurological and medical disorders.

Methods

Conceptualization and Initial Development of the Survey

The Essential Tremor Enfeeblement Survey (ETES) was developed to measure enfeeblement in individuals with ET, which we defined as a perceived debilitation or loss of vitality in patients secondary to their disease process. To our knowledge, the concept of enfeeblement had not previously been defined as a clinical phenomenon. To verify this, we performed a literature search in PubMed and PsycINFO databases. Search terms used were “enfeeblement,” with 15 and 44 results, respectively; “enfeebled,” with 48 and 38 results, respectively; and “enfeeble,” with 4 and 5 results, respectively. The contexts in which these terms were used was limited to archaic medical terminology (i.e., mental enfeeblement as it relates to psychiatric conditions such as dementia)^21-23 or an undefined or nonspecific term or descriptor (e.g., “psychosomatic enfeeblement,”²⁴ “enfeeblement of the response,”²⁵ “enfeebled hematopoiesis”²⁶); one qualitative study identified “enfeeblement of dignity” as a category comprised of 8 themes related to human dignity.²⁷ However, we found no pre-existing scales for enfeeblement. We also studied existing frailty scales. Despite the a priori conceptual similarities between frailty and enfeeblement, and the multiplicity of unique frailty scales,²⁸ we were unable to identify a measure of frailty that accurately represented and fully encapsulated enfeeblement.

Once we determined enfeeblement was a novel concept in the literature, we formulated 8 questions that would ultimately constitute the ETES (Table 1). A number of these questions were formulated based on conversations we had had with patients having ET and their caregivers. Face validity for these questions was established by expert opinion from both a senior movement disorders neurologist with a special interest in tremor and experience in scale development (EDL), and a research psychologist specializing in aging populations (JKM). Survey answers are recorded on a 5-point Likert-type scale corresponding to the level of agreement with each question (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, 5 = strongly agree), and the summed score is proportional to the degree of enfeeblement (range: 8-40) where higher score indicates more enfeeblement.

Table 1.

Essential Tremor Enfeeblement Survey.

The survey is to be completed by a caregiver. Score each item according to the scale:  1 = Strongly disagree  2 = Disagree  3 = Neither agree nor disagree  4 = Agree  5 = Strongly agree
Question	Score
1. The tremor makes your loved one suddenly seem prematurely old.
2. Because of their tremor, your loved one seems totally unable to effectively use their hands.
3. Because of their tremor, your loved one seems too disabled to undergo usual living processes.
4. Your loved one seems to be rendered helpless by their affliction.
5. Because of their tremor, your loved one seems broken down.
6. Your loved one now seems to be fading away.
7. The tremor makes your loved one seem enfeebled.
8. Watching your loved one struggle with their tremor sometimes makes you feel exhausted.
Total score

The ETES is a clinical outcome assessment, which is defined by the Food and Drug Administration as a means to “measure a patient’s symptoms, overall mental state, or the effects of a disease or condition on how the patient functions.”²⁹ Also, the ETES is an observer-reported outcome as opposed to patient reported or clinician reported. It requires close familiarity with the patient, their functional capabilities, and their general day-to-day livelihood. As patients with ET can be cognitively impaired, an observer-reported outcome was preferable.

Study Cohort

Participants were recruited through a larger, nationwide, ongoing longitudinal study of cognitive function in ET (Clinical Pathological Study of Cognitive Impairment in Essential Tremor [COGNET], NINDS R01NS086736). Eligibility criteria for COGNET were (1) ET diagnosis, (2) age ≥55 years old, (3) no history of deep brain stimulation surgery, and (4) willingness to perform neuropsychological tests and be a brain donor.³⁰ Participants undergo in-person evaluations at 18-month intervals, with tremor assessed by videotaped neurological examination to assign a total tremor score (range: 0-36, higher score indicating more severe tremor)^31-33 and using a 10-item tremor disability questionnaire (range: 0-100, higher score indicating greater disability).^34,35 Neuropsychological tests performed during the visit include Mini-Mental State Examination (MMSE, scores <24 considered abnormal),^36,37 Montreal Cognitive Assessment (MoCA, scores <26 considered abnormal),³⁸ Geriatric Depression Scale (GDS; range: 0-30, higher scores indicating greater depression),³⁹ and Generalized Anxiety Disorder 7-Item Scale (GAD-7; range: 0-21, higher scores indicating greater anxiety).⁴⁰ Participants are asked to identify someone with whom they have a close relationship and who could provide another perspective on their well-being. These informants complement participant information by completing the 8-Item Lawton Instrumental Activities of Daily Living Scale (range: 0-8, lower scores indicating more disability),⁴¹ and the Clinical Dementia Rating (0 = no dementia, 0.5 = questionable dementia, 1 = mild dementia, 2 = moderate dementia, and 3 = severe dementia).^42,43

Participants were recruited for this enfeeblement substudy by study personnel (MAZ), who contacted COGNET participants by telephone between regular study intervals (November 2017 to March 2018), inviting them to complete an online survey about their quality of life. Interested individuals were consented and asked to identify a caregiver, that is, a family member or friend who helped them with daily tasks, provided emotional support, or could give insight into their well-being. We then contacted these caregivers, consented them to participate in the study, and invited them to complete a caregiver version of the survey. Participants from the COGNET cohort were contacted in alphabetical order until 100 individuals with ET and their caregivers were enrolled, totaling 100 dyads. As suggested by Burton and Mazerolle,⁴⁴ we sought to power our analysis by enrolling at least 10 caregivers for each of the 8 questions on the ETES (i.e., n ≥ 80). The institutional review board of Yale University approved this study.

Survey Distribution and Implementation

Surveys were distributed by e-mail to individuals with ET and to their caregivers using the Yale Qualtrics online survey platform. Surveys were coded such that responders could not proceed to the next section without answering all previous questions, and responses are saved at every step, thus allowing participants to pick up where they left off if they did not initially finish the survey, and ensuring completeness of all responses. Due to the sensitive nature of the questions, participants were instructed to answer the survey independently from and without the influence of their counterpart in the dyad. The dyad comprised the individuals with ET (henceforth the care-recipients) and their corresponding caregiver.

Slightly different surveys were distributed to care-recipients and caregivers. Both were asked basic demographic information and open-ended questions about how ET has affected the dynamic of their relationship. Care-recipients completed a modified Essential Tremor Embarrassment Assessment (range: 14-70, higher scores indicating greater embarrassment).⁴⁵ Meanwhile, caregivers completed the ETES and the perceived Essential Tremor Embarrassment Assessment (range: 14-70) to assess their care-recipient’s tremor-related embarrassment,²⁰ and the Zarit Burden Inventory–Short Form (ZBI-12; range: 0-48) to assess levels of caregiver burden reported by caregivers.¹⁹ Additionally, a subset of caregivers were asked to repeat their ETES evaluation 1 to 3 months after completing the first survey; by the end of the study period, 40 responses were received. Survey data were further complemented by the regularly collected COGNET variables.

Statistical Analysis

All analyses were conducted using SPSS version 24.0 (SPSS Inc, Chicago, Illinois), and results were considered significant at P < 0.05. We performed a structured survey analysis on the ETES to establish validation beyond face validity, including survey acceptability, test–retest reliability, component analysis, and convergent validity. A subanalysis examining convergent validity among the 14 more enfeebled participants (i.e., responders who answered “Agree” or “Strongly Agree” to at least 3 of 8 items of the survey) was also performed.

Acceptability.

To determine survey acceptability, we assessed the extent of both floor and ceiling effects, with ideal cutoffs set below 15% for each,⁴⁶ meaning neither the minimum enfeeblement score (i.e., 8) nor the maximum (i.e., 40) should ideally be endorsed by more than 15% of the cohort. Also, we examined the difference between the mean and the median scores, defining an acceptable difference as ≤10% of the maximum possible score (i.e., ≤10% of 40).^45,47

Test–retest reliability.

Test–retest reliability was assessed using a 1-way random effects analysis of variance model to obtain an intraclass correlation coefficient (ICC) for each item individually as well as for the total enfeeblement score. Levels of agreement were reported as poor if ICC was <0.40, fair if ICC was 0.40-0.59, good if ICC was 0.60-0.75, and excellent if ICC was >0.75.⁴⁸

Component analysis.

Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett test of sphericity were used to determine adequacy of the ETES for component analysis. Afterward, survey items underwent principal component analysis using Varimax rotation and Kaiser normalization, and components were extracted based on both scree plot inspection and eigenvalues >1.⁴⁹ Internal consistency was determined using Cronbach α coefficient, with a score >0.70 considered adequate.⁵⁰

Convergent Validity.

To assess convergent validity, we used Spearman correlation coefficients to assess the correlations between ETES score and variables hypothesized to be correlated with enfeeblement. These included care-recipient variables (age, total tremor score, tremor disability score, MoCA, MMSE, GAD-7, and GDS score) and caregiver-associated variables (perceived embarrassment and caregiver burden). Spearman correlation coefficients were considered negligible if <0.19, weak if 0.20-0.34, moderate if 0.35-0.50, or close correlations if >.50.⁴⁷

Results

Survey Respondents

We initially recruited one hundred care-recipient/caregiver dyads, but 2 dyads were excluded: in one case, the care-recipient completed both surveys, and in the other, the identified caregiver was a paid health-care worker. All 98 remaining dyads completed their surveys in full, and their data were used for the analyses, with demographic data described in Table 2. Overall, care-recipients were older than caregivers. Most caregivers were the spouse of the care-recipient (62.2%) followed by children (22.4%). In all, 70.4% endorsed living with their care-recipient.

Table 2.

Care-Recipient/Caregiver Dyad Characteristics.^a

Characteristics	Data
Care-recipients (n = 98)
Age (years)	80.5 (±9.0)
Female	54 (55.1%)
Education (years)	16 (IQR 14-18)
MMSE score	29 (IQR 22.3-30)
MoCA score	24 (IQR 19-28)
GDS score	5 (IQR 2-9)
GAD-7 score	1 (IQR 0-3)
Total tremor score	20.7 (±5.6)
Tremor disability score	70 (IQR 45-85)
IADL (per caregiver)	8 (IQR 7-8)
CDR
0 (No dementia)	78 (79.6%)
0.5 (Questionable)	15 (15.3%)
1 (Mild)	2 (2.0%)
2 (Moderate)	3 (3.1%)
Essential Tremor Embarrassment	38.0 (± 12.3)
Assessment score
ETES total score	12 (IQR 9-18)
Caregivers (n = 98)
Age	73 (IQR 64-79)
Female	63 (64.3%)
Relationship to care-recipient
Spouse	61 (62.2%)
Son/daughter	22 (22.4%)
Brother/sister	2 (2.0%)
Other family member	3 (3.1%)
Friend	10 (10.2%)
Lives with care-recipient	69 (70.4%)
Education level
High school or equivalent (eg, GED)	16 (16.3%)
Trade school	7 (7.1%)
Associate degree	14 (14.3%)
Bachelor’s degree	34 (35.4%)
Master’s degree	21 (21.4%)
Doctorate or equivalent	6 (6.1%)
ZBI-12 score	3 (IQR 0-7)

Abbreviations: CDR, Clinical Dementia Rating; ETES, Essential Tremor Enfeeblement Scale; GAD-7, Generalized Anxiety Disorder 7-Item Scale; GDS, Geriatric Depression Scale; IADL, Instrumental Activities of Daily Living Scale; IQR, interquartile range; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment; ZBI-12, Zarit Burden Inventory–Short Form.

^aFrequencies presented with percentages. Normally distributed variables as assessed by Kolmogorov-Smirnov tests are presented with means and standard deviations; non-normally distributed variables are presented with medians and interquartile range.

Essential Tremor Enfeeblement Survey Analysis

Acceptability.

Table 3 demonstrates how caregivers responded to each ETES question. In all, 15 (15.3%) caregivers endorsed the minimum enfeeblement score; the highest recorded score was 32 of a potential maximum score of 40. The mean ETES score was 14.2 points and the median score was 12 points.

Table 3.

Caregiver Responses to Each ETES Question.

Question	1 = Strongly Disagree	2 = Disagree	3 = Neither Agree nor Disagree	4 = Agree	5 = Strongly Agree
1. The tremor makes your loved one suddenly seem prematurely old.	27 (27.6%)	35 (35.7%)	20 (20.4%)	16 (16.3%)	0 (0.0%)
2. Because of their tremor, your loved seems totally unable to effectively use their hands.	39 (39.8%)	39 (39.8%)	8 (8.2%)	11 (11.2%)	1 (1.0%)
3. Because of their tremor, your loved one seems too disabled to undergo usual living processes.	56 (57.1%)	28 (28.6%)	6 (6.1%)	8 (8.2%)	0 (0.0%)
4. Your loved one seems to be rendered helpless by their affliction.	64 (65.3%)	21 (21.4%)	9 (9.2%)	4 (4.1%)	0 (0.0%)
5. Because of their tremor, your loved one seems broken down.	59 (60.2%)	21 (21.4%)	10 (10.2%)	8 (8.2%)	0 (0.0%)
6. Your loved one now seems to be fading away.	61 (62.2%)	27 (27.6%)	4 (4.0%)	6 (6.1%)	0 (0.0%)
7. The tremor makes your loved one seem enfeebled.	53 (53.5%)	27 (27.6%)	11 (11.2%)	7 (7.1%)	0 (0.0%)
8. Watching your loved one struggle with their tremor sometimes makes you feel exhausted.	47 (48.0%)	29 (29.6%)	9 (9.2%)	12 (12.2%)	1 (1.0%)

Abbreviation: ETES, Essential Tremor Enfeeblement Scale.

Test–retest reliability.

In all, 40 caregivers repeated the ETES by the end of the study period, and test–retest reliability was assessed using these responses. Questions 2, 4, 6, and 7 demonstrated ICCs between 0.40 and 0.59, and questions 1, 3, 5, and 8 demonstrated ICCs between 0.60 and 0.75; all ICCs were associated with P < 0.001. The total ETES scores demonstrated an ICC = 0.73 (P < 0.001).

Component analysis.

The Kaiser-Meyer-Olkin measure of sampling adequacy was elevated (0.915), and Bartlett test of sphericity was significant (P < 0.001), thus demonstrating adequate conditions for principal component analysis. Scree plot inspection suggested the presence of 1 component, and principal component analysis identified only 1 component with eigenvalue >1. This one component explained 66.9% of the variance observed. As shown on the pattern matrix (Table 4), question 5 had the highest component loading, indicating the highest correlation with the extracted component; question 1 had the lowest component loading. As only 1 component was extracted after component analysis, internal consistency was measured between all 8 items of the ETES. This resulted in a Cronbach α of .92.

Table 4.

Pattern Matrix of the ETES.

Question	Component Loading
1. The tremor makes your loved one suddenly seem prematurely old.	0.624
2. Because of their tremor, your loved seems totally unable to effectively use their hands.	0.774
3. Because of their tremor, your loved one seems too disabled to undergo usual living processes.	0.856
4. Your loved one seems to be rendered helpless by their affliction.	0.871
5. Because of their tremor, your loved one seems broken down.	0.909
6. Your loved one now seems to be fading away.	0.800
7. The tremor makes your loved one seem enfeebled.	0.856
8. Watching your loved one struggle with their tremor sometimes makes you feel exhausted.	0.821

Abbreviation: ETES, Essential Tremor Enfeeblement Scale.

Convergent validity.

The ETES was weakly correlated with care-recipient age (r = 0.30, P = 0.003) and tremor disability score (r = 0.28, P = 0.006), moderately correlated with the care-recipient’s total tremor score (r = 0.40, P < 0.001) and the care-recipient’s GDS score (r = 0.45, P < 0.001), and closely correlated with the perceived Essential Tremor Embarrassment Assessment (r = 0.60, P < 0.001) and caregiver burden as measured by the ZBI-12 (r = 0.63, P < 0.001). The ETES was weakly correlated with MoCA score (r = −0.21, P =0.04) but not MMSE score (r = −0.13, P = 0.22); GAD-7 score was not significantly correlated (r = −0.08, P = 0.48).

Fourteen participants were more enfeebled (ie, answered “Agree” or “Strongly Agree” to at least 3 of 8 items of the survey), with median ETES score = 26. Among variables that were previously closely correlated with enfeeblement, perceived embarrassment remained significant (r = 0.65, P = 0.012), but ZBI-12 did not (r = 0.37, P = 0.196). Remaining variables did not approach significance.

Discussion

The ETES is a clinical outcome assessment scored by an observer who can provide insight into a patient’s enfeeblement. Its personal nature requires close familiarity to the person with ET, thus making caregivers ideal persons to answer the survey.

Responses to the ETES showed a tendency toward the lower end of possible scores (minimum 8 and maximum 40), with the median enfeeblement score being 12. Taken individually, a number of the questions showed substantial floor effect, with one-half or more of the caregivers endorsing the minimal score. When we consider the total ETES score, however, the floor effect was reduced to 15.3%, which is close to the suggested upper limit of 15%.⁴⁶ Ceiling effect for the separate items was negligible and nonexistent for the total ETES score. The absolute difference between the median and the mean enfeeblement score was <10% of the maximum score, and so by this measure, the ETES met criteria for acceptability.^45,47

Test–retest reliability of the items in the ETES was fair to good, and the sum score demonstrated an ICC = 0.73 (P < 0.001), consistent with an overall good agreement.⁴⁸

According to the principal component analysis, approximately 66.9% of the variance observed in the ETES can be accounted for by a single component, and no other components could adequately be extracted (eigenvalues <1).⁴⁹ As such, internal consistency between all questions was calculated, resulting in a Cronbach α coefficient of 0.92, which is well above the 0.70 threshold.⁵⁰ Although Cronbach a coefficient scores above 0.90 may raise concern for redundancy between questions,^50,51 we believe that at face value each question offers a unique perspective of enfeeblement in ET. Furthermore, it is worth noting that keeping all 8 questions of the ETES helps maintain the scale’s discriminant capabilities by minimizing the observed floor effect. All questions demonstrated a relatively high component loading (Table 4) and therefore correlated with the extracted component (i.e., enfeeblement).

In assessing convergent validity, we assumed the ETES would be correlated with age due to the natural loss of vitality seen with aging. Indeed, the ETES was weakly correlated (r = 0.30, P = 0.003) with care-recipient age, thus supporting our assumption. It also suggests there is more to enfeeblement than aging as it was not perfectly correlated (eg, r = 1.0). Since enfeeblement arises due to a more pervasive disease process, it is fitting that a moderate correlation was found between the ETES and the total tremor score (r = 0.40, P < 0.001) and tremor disability score (r = 0.28, P = 0.006); depressive symptoms were also moderately correlated with enfeeblement (r = 0.45, P < 0.001). The highest correlation seen with ETES was found in care-recipient embarrassment as perceived by the caregiver (r = 0.60, P < 0.001) and in caregiver burden (r = 0.63, P < 0.001); much like the ETES, both variables are dependent on the caregiver. Our subanalysis exploring these same correlates among those who were more enfeebled did not yield significant results outside perceived embarrassment (r = 0.647, P = 0.012), but this may be due to the small sample size (n = 14). Still, we were overall satisfied with the convergent validity of the ETES.

Although the ETES shows favorable results on survey validation tests, these must be taken within the contexts of the study’s limitations. First, the ETES was piloted using a sample of a larger cohort of interested and highly motivated brain donors participating in the COGNET study; we recognize that this sample may not necessarily be representative of all patients with ET. Moreover, although items on the ETES were independently endorsed by caregivers when answering open-ended questions about their care-recipient/caregiver dynamic, neither care-recipients nor caregivers were consulted during creation of the ETES, which would have otherwise helped with item generation, optimal wording, and overall comprehension, thus bolstering content validity of the survey.^29,52 Finally, enfeeblement scores in our sample were mostly mild, and we were unable to reproduce the same correlations among the smaller group of high scorers; although this may be due to lack of power, it does highlight the need for a larger, more enfeebled cohort to be analyzed.

Initial results of the ETES are encouraging based on our validation assessments and warrant further exploration to determine the role of enfeeblement as a clinical outcome measure. Future steps include exploring enfeeblement’s association with caregiver burden, studying enfeeblement in other tremor disorders and in cohorts which may be more enfeebled, eventually adapting current enfeeblement items to other nontremor disease processes, assessing what impact enfeeblement may have within patient outcome models, and ultimately determining how this may influence clinical decision-making.