The Facial Disability Index: reliability and validity of a disability assessment instrument for disorders of the facial neuromuscular system.

[VanSwearingen JM, Brach JS. The Facial Disabilityy Index: reliability and validityy of a functional assessment instrument for disorders of the facial neuromuscular system. Phys Ther. 1996;76:1288-1300.]

Key Words: Facial paralysis; Tests and measurements, functional.

The facial motor system is responsible for functions critical for physical, social, and psychological well-being. Damage to the facial motor system includes those conditions affecting facial neurons in the brain stem, the facial nerve and its nerve branches, and the facial muscles. These conditions can result in deficits in eating, drinking, speaking, conveying conversational signals (eg, punctuation signs conveyed by movements of the eyebrows during speech), and even conveying intimate human information (eg, anger, disgust, happiness, surprise).[1]

Facial palsy and altered facial movements can result in marked disfigurement of the face at rest as well as in posed (voluntary) or spontaneous expressions. The impact that such facial disfigurement has on social interactions has been emphasized by MacGregor[2] in his studies of social and psychological consequences of facial deformities. Particular difficulties involve reactive responses in social communications and psychological and physical fatigue related to managing social interactions.[2, 3]

Patients with facial neuromuscular dysfunction have disability associated with the disorder.[4] The terms "impairment" and "disability" are used in this report as defined within the World Health Organization's International Classification of Impairments, Disabilities, and Handicaps[5] disablement scheme. Impairments refer to any physiologic or anatomic abnormalities at the organ or tissue system level, such as muscle weakness. Disabilities are person-level problems characterized by the inability to perform any of the activities considered usual for a human being, such as limitations in walking or limited ability to communicate. The illness model of Mechanic and Volkhart[6] and the health-related quality-of-life (HRQL) model recently proposed by Wilson and Cleary[7] suggest that the patient's emotional and social well-being may influence the relationship between impairments and disabilities. Thus, as Brook and Kamberg[8] suggested in a discussion of general health status measures and outcome measurement, measuring change in disease specific physiologic measures (impairment domain) alone as the outcome of clinical interventions is insufficient.

Motion analysis[9-11] and observer-based rating scales[12-14] for quantifying impairments of facial resting posture and movement have been described for patients with facial neuromuscular dysfunction. Measures of limitations in physical disability (eg, problems with basic and instrumental activities of daily living and difficulty with producing appropriate facial expressions and appearance) of individuals with facial nerve disorders have not been developed. What is known about the physical disabilities associated with facial nerve disorders comes from the comments of patients,[14,15] but systematic measurement of the limitations of a person's usual physical activities associated with facial nerve disorders is lacking.

Psychological and social difficulties experienced by patients with facial neuromuscular dysfunction and acknowledged by clinicians and psychologists range from altered emotional well-being, decreased selfesteem, anxiety, depression, and alternative behaviors such as social isolation and addiction.[1,3,4,14] Problems related to personal and work relations, involvement in social activities, and even secondary illness stemming from facial dysfunction are common.[3,14-16] Despite the clinical recognition of such problems, no report of the systematic measurement of psychological and social factors related to facial neuromuscular dysfunction could be identified in an extensive review of the literature.

To enhance the assessment of facial neuromuscular dysfunction beyond the impairment domain, we developed a disability assessment instrument, the Facial Disability Index (FDI) (Appendix). The FDI is a self-report, disease-specific instrument designed to provide the clinician with information about the disability and related social and emotional well-being of patients with facial nerve disorders. The purposes of this investigation were to examine the reliability and construct validity of the FDI and to determine the usefulness of the FDI compared with a more general HRQL instrument in the clinical assessment of facial nerve disorders.

Method

The FDI is a brief, self-report questionnaire of physical disability and psychosocial factors related to facial neuromuscular function. Standard disease-specific HRQL measures (eg, Oswestry Low Back Pain Disability Questionnaire,[17] Neck Disability Index[18]) and general HRQL measures (eg, Medical Outcomes Survey Health Status Questionnaire [SF-36],[19] Functional Status Questionnaire [FSQ][20]) with an emphasis on mobility and functions involving the extremities appeared to be inadequate for assessing the specific problems of individuals with facial movement disorders. The FDI was developed by clinical researchers from the Facial Nerve Center, University of Pittsburgh Medical Center (UPMC), to provide an account of the patient s daily experience of living with a facial nerve disorder. The FDI was intended to assess disability and the outcome of intervention in terms of meaningful change in the patient s physical disability and psychosocial status.

The format and scoring of the questionnaire items were based directly on the FSQ,[20] a comprehensive, selfreport questionnaire assessing multiple domains of health used for screening disability and monitoring clinical change in the functional disability status of ambulatory patients. As a disease-specific disability measure, the range of disability assessed by the FDI was narrowed to the domains of physical function and social/well-being function (including psychological and social/role function). The social/well-being function subscale of the FDI includes content of some items drawn from the psychological and social/role function domains of the SF-36,[19] the FSQ,[20] and the Beck Depression Inventory,[21] adapted for the assessment of facial nerve disorders. In an effort to enhance the content validity of the instrument,[20] questions were selected and adapted from the disability component of other instruments based on patients recorded complaints, describing the problems they experienced in usual human functions and roles living with a facial nerve disorder.[14,15] The initial form of the FDI was refined slightly after the first 6 months of use, based on commonly noted physical disabilities that patients described to the physical therapists at the Facial Nerve Center. For example, difficulty brushing the teeth or rinsing the mouth was frequently identified as a problem, and an of ten recognized sign of recovery was the improved ability to brush the teeth or avoid leaking fluid when rinsing the mouth.

The format and scoring of the FSQ were chosen because of (1) the flexibility in obtaining summary scores for subscales regardless of the number of questions answered, (2) the ability to allow the patient to distinguish disability due to facial neuromuscular dysfunction from perhaps similar difficulties caused by an alternative disorder, (3) the readily understandable subscale score meaning (de, scores transformed to a 100-point basis), and (4) the brevity of the questionnaire.[20] The FSQ had been designed and implemented for an ambulatory outpatient population, similar to the status of the majority of the population with facial nerve disorders seen in the clinic and expected to complete the FDI.

In our clinical setting, outcomes of facial rehabilitation are routinely assessed using (1) physical performance measures of impairments of facial movement and (2) a self-report of physical function and social/well-being function (FDI). Patients who agree to be evaluated in the Facial Nerve Center complete these assessments. Individuals who elect to participate in facial rehabilitation are assessed prior to and at regular intervals after the onset of facial rehabilitation.

An additional battery of standardized measures of psychological and social function are also administered to patients who provide written informed consent to participate in a Facial Nerve Center study of psychosocial aspects of facial nerve disorders. The routine clinical assessments of physical impairments, disability, and HRQL and the study of psychosocial factors of individuals with facial neuromuscular disorders have been approved by the Institutional Review Board for Biomedical Research of the University of Pittsburgh.

Sample

Between January 12, 1995, and May 9, 1995, we administered the FDI to 46 ambulatory patients (30 female, 16 male) of the Facial Nerve Center, UPMC. Patients were included in the sample if they (1) had a history of a disorder of the facial neuromotor system and some residual facial neuromuscular dysfunction, (2) demonstrated or reported some difficulty on one or more basic or instrumental activities of daily living or on social activities included in the FDI, (3) were age 18 years or older, (4) could speak and read English, and (5) had no apparent disorientation to time and place. All patients completed the assessments of facial neuromuscular dysfunction. The patients' mean age was 46.8 years (SD=15.6), with a range of 21 to 80 years. Approximately 52% of the assessments were initial evaluations; the remaining 48% were reassessments of patients recovering following surgical facial reanimation or participating in facial rehabilitation for recovery of facial neuromuscular function.

Reliability

The degree to which scores reflect real phenomena rather than chance variations is reliability. For any clinical measure to be useful in assessing a patient's problem and the effectiveness of interventions necessitates using a measure with high reliability.[22] The assessment of reliability and the improvement of reliability for composite measures pose several technical difficulties. According to Armor,[23] the most commonly applied procedure for determining internal-consistency reliability of composite measures, Cronbach's alpha, may not produce optimum alpha reliability because composite measures often do not meet the mathematical assumptions of alpha reliability, nor does alpha reliability include a standard process of accounting for multidimensionality (eg, subsets of mutually independent items that cluster within the composite).[23] We examined the internal consistency reliability of the two subscales of the FDI (physical function and social/well-being function) using theta reliability, as described by Armor[23] and based on principal-component factor analysis (PCFA), to assess optimal reliability.

Principal-component factor analysis is a statistical method of identifying the number of dimensions within a multivariate set of items and indicating the contribution of each item to a given dimension. A single item within a multivariate scale can contribute differently to different dimensions, called the "factors" of factor analysis. Principal-component factor analysis offers an alternative to other measures of composite reliability based on traditional parallel-forms assumptions and inter-item correlations (eg, Cronbach's alpha). The traditional composite reliability methods for multivariate scales assume that all items within a composite measure have similar inter-item correlations and that each item contributes similarly to a single scale dimension. When the responses to items of a scale are prone to be influenced by attitudes or behavioral attributes, as may be true of the FDI, individual scale items often contribute to a single factor unequally, or the items contribute equally or unequally to two or more factors of the scale. For such scales, the mathematical assumptions of alpha reliability are violated, and optimal reliability is compromised.[23]

Using PCFA, a subject's score on any item is broken down into numerous factors, termed "factor scores." The contribution of each item to a given factor determines the weighting of the item for that factor, represented by the factor loadings. Thus, each factor explains a certain amount of the variance of the construct being measured (eg, disability associated with facial neuromuscular dysfunction), and each factor represents a statistically independent, differential contribution of a set of individual items to the construct.[23]

The theta reliability coefficient, derived from the PCFA, is an estimate of composite reliability based on the factor scores. In a sense, the factor score, or the score of a subject on a given factor, is a weighted sum of individual items using the factor loadings as weights, and thus is representative of the composite scale score. In PCFA, the amount of variance accounted for by a particular factor is known as the latent root (also referred to as the eigenvalue or characteristic root). The formula for theta reliability used in this study is: [p / (p --1)] [(L -- 1) / L], where p is the number of items in the scale and L is the latent root of the factor representing the component loadings of the subscale items of the PCFA, with rotation.[23,24] When the latent root for the second factor is greater than 1 and a number of scale items have higher loadings on the second factor than on the first factor, a two-factor solution is indicated and the rotated loadings are used. Theta coefficients for composite reliability range from 0 to 1, with 1 indicating perfect reliability.[23]

The theta coefficient provides optimal reliability for composite scales when the individual items differentially contribute to the overall construct. Theta and alpha reliability coefficients would be equal if each item contributed similarly to the construct and the inter-item correlations were close to the average correlation for the scale items. When the individual items contribute differently to the covariance, weighting items based on the average amount of variance due to an item (expressed as the factor loadings) results in a composite scale that more accurately demonstrates the ratio of covariation to the scale variance. Theta coefficients are usually greater than or equal to the alpha coefficient of reliability.[23]

Principal-component factor analysis was chosen because we felt the scale items would not equally contribute to the overall construct of facial neuromuscular dysfunction. Principal-component factor analysis also provided a means of factor scaling (the identification of sets of items with a higher correlation with a given factor) and the possibility of multiple-factor solutions enhancing reliability.[23,24]

Validity

To measure validity of the FDI (de, the degree to which the scale measures what it was intended to measure), construct validity was examined. Construct validation of the instrument involved comparing the scores of the FDI with clinical measures that a person might expect to be associated with facial function.[20,25] Pearson's product moment correlation coefficients or Spearman's rankorder correlation coefficients were used to determine the bivariate relationships between FDI subscale (physical function and social/well-being function) scores and other external clinical measures one might expect to be associated with facial disability. The clinical measures included physical examination and psychosocial factors.

We expected the FDI physical function subscale to be directly associated with the physical performance measure of voluntary facial movement, determined using the Facial Grading System (FGS) movement subscale.[26] The FGS is an impairment-level, criterion-based rating scale of resting posture, voluntary movement, and abnormal movement (de, synkinesis) of facial neuromuscular disorders. The FGS movement subscale score is the sum of the therapist's ratings of movements in each of five regions of the face, obtained by assessing the movements of brow raise, eye closure, open-mouth smile, snarl (eg, showing the upper teeth), and pucker. Voluntary movement is rated on a scale of 1 to 5, with 1 representing no movement and 5 meaning facial movement equal to the movement of the uninvolved side of the face. The reliability and validity of the FGS have been demonstrated previously.[26,27] We suspected that individuals with better facial movement would have less physical disability.

We hypothesized that the FDI social/well-being function subscale scores would not be related to the facial movement impairment measure, but would be directly associated with measures of psychological and social wellbeing. We expected that individuals with higher FDI social/well-being function subscale scores would be more likely to be less psychologically distressed by their facial neuromuscular dysfunction. The hypothesis was tested by determining the clinical status of psychological distress for the subset of the sample (n=14) who had agreed to participate in a concurrent Facial Nerve Center study of psychosocial aspects of facial nerve disorders and completed the battery of standardized assessments of psychological and social function. The clinical status of psychological distress was determined using the Primary Care Evaluation of Mental Disorders (Prime MD),[28] a screening tool for detecting mood disorders in a primary care population. The Prime-MD involves a self-report patient questionnaire and a follow-up clinical interview assessment of specific mood-disorder symptoms triggered by responses to the questionnaire.[28] Using scores on the questionnaire and the interview, a clinical status of psychological distress was categorized as (1) positive diagnosis of depression, anxiety, or increased bodily awareness (somatoform disorder), (2) screened positive but not diagnostic for psychological distress, or (3) no symptoms of psychological distress, and given a psychosocial status code (0, 1, or 2, respectively).

Because the FDI physical function and social/well-being function subscales are both assessments of person-level function, we suspected that the two subscales would be modestly, but directly, related to each other. We expected that patients would reduce their interaction with others if they perceive that their physical disability is severe enough to result in frequent incidences of losing food or leaking fluids from the mouth when eating or drinking.

We also assessed the construct validity of the subscales of the FDI by comparing the relationships of the FDI subscale scores with the relationships of the total FDI scale and the other clinical measures. If the subscales of the FDI identified by PCFA are the better descriptors of the patients' disability[5] associated with a facial nerve disorder, the expected relationship between the subscale scores and clinical measures should be enhanced over the same relationship for a single FDI scale score.[23]

Lastly, we were interested in determining whether the disease-specific FDI provided a better indication of the problems of the individual with a facial neuromuscular disorder than did a more general HRQL instrument, such as the SF-36. The SF-36 is a standardized self-report questionnaire of HRQL in eight major medical outcome dimensions: physical functioning, physical role function, emotional role function, social functioning, mental health, energy and fatigue, bodily pain, and general health.[19] The SF-36 has been used extensively.[29,30] For the subset of the sample who agreed to participate in a concurrent study of psychosocial aspects of facial nerve disorders and completed the SF-36 (n=18), a Pearson product-moment correlation was used to determine the relationship between the FDI and the related subscales of the SF-36 with the other clinical measures of facial nerve disorders.

If the SF-36 was adequate for describing the problems of individuals with facial neuromuscular disorders, we would expect to find a direct relationship between the subscales of the SF-36 and the clinical measures of physical performance and psychological distress, similar to the relationships demonstrated for the FDI subscales. Specifically, we hypothesized that the SF-36 physical functioning subscale and the facial movement measure would not be directly related, or a weaker relationship would be demonstrated for the SF-36 physical functioning than for the FDI physical function subscale and facial movement for the subset of patients with facial neuromuscular dysfunction.

The FDI social/well-being function subscale includes components assessed under the SF-36 dimensions of social functioning, emotional role function, and mental health. We expected that the SF-36 social functioning, emotional role function, and mental health subscales would be directly, but less highly, related to psychosocial status than the FDI social/well-being function subscale for the subset of patients with facial neuromuscular disorders studied.

A basic assumption of models of illness[5-7] and a primary foundation of interventions is that some underlying relationship exists among the patient's physiological impairment, disability, and social and emotional wellbeing.[16,31] We would expect the person who becomes ill and seeks treatment to be experiencing some disruption of his or her ability to perform daily activities and some disruption of feelings of well-being. Because of the obvious social consequences of a facial neuromuscular disorder associated with facial disfigurement,[2] we might expect a closer relationship among the domains of illness than would typically be observed with other less disfiguring disorders. Thus, if the SF-36 instrument is adequate for targeting the assessment of the patient's difficulties directly associated with living with a facial neuromuscular disorder, then we would expect the instrument to demonstrate a relationship between the subscale measures of physical disability and psychosocial status as well the relationships described.

Results

Analysis of internal consistency of the FDI subscale scores demonstrated acceptable reliability for a clinical instrument, using theta reliability for composite measures. Theta reliability was .88 for the FDI physical function subscale and .83 for the FDI social/well-being function subscale, based on the latent root of the factor analysis (Tab. 1). The analysis of the FDI subscale item scores for the sample confirmed that the FDI physical function subscale items form a homogenous group, clearly separable from the FDI social/well-being function subscale items (Tab. 2). Only factor loadings greater than 0.5 are highlighted in Table 2, illustrating the separability of the first-factor (de, FDI physical function subscale) and second-factor (de, FDI social/well-being function subscale) items.

Table 1.
Principal Component Confirmatory Factor Analysis of Facial
Disability
Index Self-Report of Functional Status(a)

                           First Factor   Second Factor

Latent root (eigenvalue)    3.364          2.953

Percent of total
  variance explained       33.64          29.52

(a) For simplicity, factors with eigenvalues of <1.0 are omitted.
Table 2.
Separability of the Facial Disability Index Physical Function and
Social/Well-being Function Subscale Items

                           Physical Function   Social/Well-being
                           Subscale (Factor    Function Subscale
Item                       Loadings)(a)        (Factor Loadings)(a)

1. Difficulty eating         .823                 .290
2. Difficulty drinking
   from a cup                .783                 .132
3. Difficulty speaking       .873                 .084
4. Excessive tearing or
   drying of the eye         .657                 .186
5. Difficulty brushing/
   rinsing teeth             .802                 .120
6. Feel calm and
   peaceful                  .340                 .619
7. Isolate self from
   people                    .233                 .841
8. Irritable toward
   those around you          .063                 .769
9. Wake up early,
   wake up several
   times                    -.166                 .555
10. Avoid going out to
   eat, shop, or
   participate in social
   activities                .186                 .897

(a) Factor loadings from principal-component factor analysis;
rotated loadings.

In Table 3, the bivariate correlations for the FDI subscale scores and the other clinical measures are shown. As expected, the FDI physical function subscale scores were directly correlated with the clinical measure of voluntary facial movement.

[TABULAR DATA 3 OMITTED] The FDI social/well-being subscale scores were not associated with the facial movement ratings (Tab. 3). The patients' report of physical disability (FDI physical function subscale scores) related to their facial nerve disorder was directly correlated with the self-report of social well-being (FDI social/well-being function subscale scores).

The bivariate correlations for the FDI social/well-being function subscale scores and psychosocial status for the subset of the sample (n=14) who completed the battery of psychological and social function assessments are also shown in Table 3. The FDI social/well-being subscale scores were directly correlated with the clinical assessment of psychosocial factors for this subset of the sample (Tab. 3).

The bivariate relationship between a single scale score for the FDI that includes both FDI subscale items (FDI total) and the clinical measures was less directly correlated (Tab. 4). The FDI subscale scores were more closely associated with the clinical assessment of the patients' facial nerve disorder.

The bivariate correlations of the relationship between the SF-36 subscales and the clinical measures of facial movement and psychosocial status for a subset of the sample (n=18) are presented in Table 5. As hypothesized, we did not observe a direct relationship between the SF-36 physical functioning subscale measure and facial movement, as was demonstrated for the FDI physical function subscale for the subset of patients with facial neuromuscular dysfunction. As we also expected, the SF-36 social functioning, emotional role function, and mental health subscales were all directly related to psychosocial status. Unexpectedly, we found the magnitude of these correlations to be similar to the magnitude for the relationship between the FDI social/well-being function subscale and psychosocial status.

Discussion and Conclusion

We believe that the FDI, as a disease-specific, self-report functional status instrument, provides an essential component of the assessment of individuals with disorders of the facial neuromuscular system. The information gathered through the use of functional status instruments such as the FDI plays as important a role as impairment measures in determining the quality of care provided for patients with facial nerve disorders.[32] The outcomes research "revolution" in medicine, physical therapy, and other therapies emphasizes the need to focus on the influence of care on patient function and HRQL.[4]

Relative to the framework of disablement defined by the World Health Organization's International Classification of Impairments, Disabilities, and Handicaps,[5 ]the FDI addresses the concept of disability: person-level behavior and functioning in expected roles.[33] The FDI can be used as an initial assessment tool and as a monitoring instrument, providing the clinician with the patient's view of the outcome of the intervention in progress. The format, a brief, self-report questionnaire of common behaviors, lends itself well to the typical outpatient practice setting. Patients can complete the FDI in a few minutes, with the information available for immediate comparison with previous scores over the course of treatment or entered into a clinical database for use in reviewing quality of care.

Based on the findings presented, the FDI subscales produce reliable measurements with construct validity. The reliability value for each subscale was above .80, a level considered acceptable for group comparisons and for monitoring functional disability in the same patient with self-report instruments.[20,34] The reliability values attained for the FDI subscales are similar to the reliability estimates for other functional status measures with a similar number of items and content.[20] The higher reliability of the FDI physical function subscale when compared with that of the FDI social/well-being function subscale may be due to the fact that the facial functions assessed are largely behaviors that necessarily occur more than once a day, every day. Many of these facial behaviors occur in the presence of others; thus, the reactive responses of others to the facial dysfunction continuously redefine the disability for the patient.[2] To further the application of the FDI for monitoring change in a patient's function, the stability of the FDI subscale scores when the FDI is administered at different points in time (test-retest reliability) will need to be determined.

In the absence of a generally agreed-on standard for quantifying facial function, construct validity was demonstrated by the association of the FDI physical function and social/well-being function subscales with the clinical measures of facial movement and psychosocial status, respectively. The relationship between facial movement ratings and the FDI physical function subscale scores also demonstrates the conceptual relationship between impairment and disability embodied in the disablement model,[5] and the underlying paradigm of physical therapy practice.[4]

The expected correlations were higher for the two separate subscales compared with the correlation for a single, composite scale version of the FDI, supporting the validity of the subscales format.[23] Although it might be desirable to have a single functional status score to report, summing the items for both subscales (1) produces a less valid measure of the disability, (2) will most likely reduce the reliability of either subscale, and (3) misrepresents the clinical phenomena of interest.[35]

It is not surprising that the SF-36 physical functioning subscale was not related to the clinical measure of facial movement, as the component items of the scale are all related to assessing activities of daily living involving mobility or lifting; none of the component items require facial neuromuscular function to complete the scale. The lack of a relationship between the SF-36 physical functioning subscale and facial movement--whereas the FDI physical function subscale and facial movement were related--suggests that the two subscales are measuring different aspects of physical disability. The social functioning, emotional role function, and mental health subscales of the SF-36, however, were directly and substantially related to the psychosocial status of the patients observed, much the same as the relationship between the FDI social/well-being function subscale and psychosocial status.

One interpretation of these results is that the disease specific instrument, the FDI, is representing the relationship among the domains of illness for patients with facial neuromuscular disorders, and that this relationship is not represented by the SF-36. We found that the SF-36 physical functioning subscale was not related to the FDI social/well-being function subscale score (r=.366), despite the fact that both the SF-36 physical functioning subscale and the FDI social/well-being function subscale were related to the SF-36 social functioning, emotional role function, and mental health dimensions (r=.737, .306, and .513 and r=.721,.621, and .488, respectively; P<.05). Patients with facial neuromuscular dysfunction can have, and often do have, other health problems that could affect their physical mobility, lifting-related performance, and social and emotional well-being. Thus, the relationship between the SF-36 physical and social/ emotional subscales may be related to health problems other than the disorder of facial movement. Only the disease-specific FDI instrument captures the physical and social/emotional impact of a facial nerve disorder.

Confidence in the validity of the FDI physical function and social/well-being subscales for representing the disability of patients with facial nerve disorders and favoring of the disease-specific measure over the HRQL measure for this population is limited by the small sample size, particularly the subset samples used for some of the comparisons. The population of patients with facial nerve disorders is small and only a small number of clinical centers dedicated to the treatment of facial nerve disorders exist in the United States, but the magnitude of the health-related problems of individuals with facial nerve disorders is great. The results of the assessment of disability associated with facial nerve disorders is shared in this preliminary stage of clinical testing to promote attention to describing patient outcomes and the opportunity for more widespread use of measures of disability among clinicians working with a similar patient population.

With total or regional facial paralysis, common functional tasks such as keeping food in the mouth while eating, drinking from a cup without leaking fluid, and speaking without slurring words can become difficult. Conversational signals of interest, approval, disbelief, and even punctuation may be difficult for the individual with facial palsy to share. Facial paralysis can mean the loss of the ability to convey any or all of at least six universal human facial expressions.[1] The inability to express anger, surprise, distress, disgust, fear, and happiness may severely impair a person's ability to be understood, or more importantly to not be misunderstood. Thus, important aspects of a patient's daily function such as emotional well-being, behavioral competence, sleep and rest, energy and vitality, and general life satisfaction[20] can be altered by impairments associated with facial nerve disorders. To provide important benefits through therapeutic intervention means providing benefits that the patient values.[32] The FDI subscales produce reliable measurements, with construct validity for measuring patient-focused (disability) outcomes for individuals with disorders of the facial motor system.

Table 4.
Construct Validity of Facial Disability Index (FDI):
Correlations of Subscale and Total Scale Scores With
the Other Clinical Assessments

                              Facial
                            Movement(a)      Psychosocial
                               (n)           Status(b) (n)

FDI physical function         .507              .231
                              (46)              (15)
FDI social/well-being
function                      .066              .694(c)
                              (44)              (14)
FDI total                     .330              .555
                              (44)              (14)

(a) Pearson's product-moment correlation coefficients, with
pair-wise comparisons.
(b) Spearman rank-order correlation coefficients, with pair-wise
comparisons.
(c) P<.01.
Table 5. Relationship of the Disease-Specific Facial
Disability Index (FDI) and the General SF-36 Scale With
Clinical Measures of Facial Neuromuscular Dysfunction

                              Facial
                            Movement(b)     Psychosocial
                               (n)          Statusb (n)

FDI physical function    .611(c)
                            (18)
SF-36
  Physical functioning      .244
                            (18)
FDI social/well-being
  function                                    .771(c)
                                               (14)
SF-36
  Social functioning                          .724(c)
                                              (15)
Emotional role function                       .744(c)
                                              (15)
Mental health                                 .796(c)
                                              (15)

(a) Pearson's product-moment correlation coefficients, with
pair-wise comparisons.
(b) Spearman rank order correlation coefficients, with
pair-wise comparisons.
(c) P < .01.

References

[1] Ekman P. Psychosocial aspects of facial paralysis. In: May M, ed. The Facial Nerve. New York, NY: Thieme Medical Publishers Inc; 1986:781-787. [2] MacGregor FC. Facial disfigurement: problems and management of social interaction and implications for mental health. Aesthetic Plast Surg. 1990;14:249-257. [3] Twerski AJ, Twerski B. The emotional impact of facial paralysis. In: May M, ed. The Facial Nerve. New York, NY: Thieme Medical Publishers Inc; 1986:788-794. [4] Jette AM. Outcomes research: shifting the dominant research paradigm in physical therapy. Phys Ther. 1995;75:965-970. [5] International Classification of Impairments, Disabilities, and Handicaps. Geneva, Switzerland: World Health Organization; 1980. [6] Mechanic D, Volkhart EH. Illness behaviour and medical diagnosis. Journal of Health and Human Behavior. 1960;1:86-94. [7] Wilson IB, Cleary PD. Linking clinical variables with health-related quality of life. JAMA. 1995;273:59-65. [8] Brook RH, Kamberg CJ. General health status measures and outcome measurement: a commentary on measuring functional status. J Chronic Dis. 1987;40:1315-1365. [9] Johnson PC, Brown H, Kuzon WM, et al. Simultaneous quantification of facial movements: the maximal static response assay of facial nerve function. Ann Plast Surg. 1994;32:171-179. [10] Burres SA. Facial biomechanics: the standards of normal. Laryagoscope. 1985;95:708-714. [11] Murty GE, Diver JP, Kelly PJ, etal. The Nottingham System: objective assessment of facial nerve function in the clinic. Otolaryngol Head Neck Surg. 1994;110:156-161. [12] House JW. Facial nerve grading systems. Laryngoscope. 1983;93: 1056-1069. [13] House JW. Facial nerve grading systems. In: Portman M, ed. The Facial Nerve. New York, NY: Masson Publishers; 1985:35-41. [14] Ross B, Nedzelski JM, McLean JA. Efficacy of feedback training in long-standing facial nerve paresis. Laryngoscope. 1991;101:744-750. [15] Brudny J, Hammerschlag PE, Cohen NL, et al. Electromyographic rehabilitation of facial function and introduction of a facial paralysis grading scale for hypoglossal-facial nerve anastomosis. Laryngoscope. 1988;98:405-410. [16] Jette AM. Using health-related quality-of-life measures in physical therapy outcomes research. Phys Ther. 1993;73:528-537. [17] Fairbanks JCT, Couper J, Davies JB, et al. The Oswestry low back pain disability questionnaire. Physiotherap,. 1980;66:271-273. [18] Vernon H, Mior S. The neck disability index: a study of reliability and validity. J Manipulative Physiol Ther. 1991;14:409-413. [19] Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). Med Care. 1992;30:473-483. [20] Jette AM, Davies AR, Cleary PD, et al. The functional status questionnaire: reliability and validity when used in primary care. J Cen Intern Med. 1986;1:143-149. [21] Beck AT, Ward CH, Mendelson M, Erbaugh MJ. An inventory for measuring depression. Arch Gen Psychiatry. 1961;4:561-571. [22] Kirshner B, Guyatt G. A methodological framework for assessing health indices. J Chronic Dis. 1985;38:27-36. [23] Armor DF. Theta reliability and factor scaling. In: Costner HL, ed. Sociological Methodology, 1973-1974. San Francisco, Calif:Jossey-Bass Inc Publishers; 1974:17-50. [24] Rosenthal R, Rosnow RL. Essentials of Behavioral Research. New York, NY: McGraw-Hill Book Co; 1984:419-425. [25] Deyo RA. Comparative validity of the sickness impact profile and shorter scales for functional assessment in low-back pain. Spine. 1986; 11:951-954. [26] Ross B, Fredet G, Nedzelski JM. Development of a sensitive clinical facial grading system. Otolaryugol Head Neck Surg. 1996;114:380-386. [27] Brach JS, VanSwearingen JM, Delitto A, Johnson PC. Impairment and disability in individuals with facial neuromuscular dysfunction. Presented in poster format at the Combined Sections Meeting of the American Physical Therapy Association; Reno, Nev; February 6, 1995. [28] Spitzel RL, Williams JBW, Kroenke K, et al. Utility of a new procedure for diagnosing mental disorders in primary care: The Prime-MD Study. JAMA. 1994;272:1749-1756. [29] Sturm R, Wells KB. How can care for depression become more cost-effective? JAMA. 1995;273:51-58. [30] Nerenz DR, Repasky DP, Whitehouse FW, Kahkonen DM. Ongoing assessment of health status in patients with diabetes mellitus. Med. Care. 1992;30:MS112-MS123. [31] Waddell G, Main CJ, Morris EW, et al. Chronic low-back pain, psychologic distress, and illness behavior. Spine 1984;9:209-213. [32] Donabedian A. Quality assessment and assurance: unity of purpose, diversity of means. Inquiry. 1988;25:173-192. [33] Jette AM. Diagnosis and classification by physical therapists: a special communication. Phys Ther. 1989;69:967-969. [34] Helmstadter GC. Principles of Psychological Measurement. New York, NY: Appleton-Century-Crofts; 1973:251-256. [35] Feinstein AR. Clinimetrics. New Haven, Conn: Yale University Press; 1987:35-37, 124-127.

Appendix

Facial Disability Index Please choose the most appropriate response to the following questions related to problems associated with the function of your facial muscles. For each question, consider your function during the past month:

Physical Function

1. How much difficulty did you have keeping food in your mouth, moving food around in your mouth, or getting food stuck in your cheek while eating?

Usually did with: Usually did not eat because: 5 no difficulty 1 of health 4 a little difficulty 0 of other reasons 3 some difficulty 2 much difficulty

2. How much difficulty did you have drinking from a cup?

Usually did with: Usually did not drink because:

5 no difficulty 1 of health 4 a little difficulty 0 of other reasons 3 some difficulty 2 much difficulty

3. How much difficulty did you have saying specific sounds while speaking? Usually did with Usually did not speak because:

5 no difficulty 1 of health 4 a little difficulty 0 of other reasons 3 some difficulty 2 much difficulty, slurring most of speech

4. How much difficulty did you have with your eye tearing excessively or becoming dry? Usually had: Usually did not have tearing because:

5 no difficulty 1 of health 4 a little difficulty 0 of other reasons 3 some difficulty 2 much difficulty

5. How much difficulty did you have with brushing your teeth or rising your mouth?

Usually did with Usually did not have difficulty

brushing or rinsing because:

5 no difficulty 1 of health 4 a little difficulty 0 of other reasons 3 some difficulty 2 much difficulty

Social/Well-being Function

6. How much of the time have you felt calm and peaceful?

6 all of the time 5 most of the time 4 a good bit of the time 3 some of the time 2 a little bit of the time 1 none of the time

7. How much of the time did you isolate yourself from people around you?

1 all of the time 2 most of the time 3 a good bit of the time 4 some oft he time 5 a little bit of the time 6 none of the time

8. How much of the time did you get irritable toward those around you?

1 all of the time 2 most of the time 3 a good bit of the time 4 some oft he time 5 a little bit of the time 6 none of the time

9. How often did you wake up early or wake up several times during your nighttime sleep?

1 every night 2 most nights 2 a good number of nights 4 some nights 5 a few nights 6 no nights

10. How often has your facial function kept you from going out to eat, shop, or participate in family or social activities?

1 all of the time 2 most of the time 3 a good bit of the time 4 some of the time 5 a little bit of the time 6 none of the time

Scoring: Physical Function Social/Well-being Function

Total Score (questions 1-5) - N/N x 100/4 Total Score (questions 6-10) - N/N x 100/5

N = number of questions answered

Invited Commentary

VanSwearingen and Brach are to be commended for their pioneering work in creating the first facial disability measure. Clearly, this report presents the groundwork on which subsequent validation studies will be built. My comments address the challenges related to developing a disability index.

Traditionally, generic and disease- or condition-specific measures have been conceived with group decision making in mind.[1-4] For example, developers have been interested in obtaining information about the health status of a group at a given point in time, comparing health status between groups at a given point in time, and assessing change in health status of a group over time. More recently, there has been a growing interest in applying disability measures to aid in decision making about individual patients. The Facial Disability Index (FDI) is a condition-specific measure that is intended to be used to aid clinicians in making decisions about individual patients. As such, it will be required to distinguish among individuals with varying levels of disability and to assess change within a person over time.

The article by VanSwearingen and Brach addresses the initial stage of instrument development. Topics include item generation, item scoring, scale development, and validation. Items for the FDI were based on existing generic measures and condition-specific measures targeted at other conditions. An important aspect of the instrument development consisted of involving patients in the item-generation phase. In this article, 10 items are reported, and it would be interesting to know the steps that were taken to come up with these items (de, Were more items analyzed formally?). The 10 items were subjected to factor analysis, and two distinct factors emerged. The factors or subscales were labeled "physical function" and "social/well-being function." Each subscale has five items, and each item has six response options. It would be informative to know the item means and standard deviations. This information would provide insights into item difficulty, furnish subsequent investigators with a benchmark, and provide researchers who are interested in using the tool as an outcome measure in clinical trials with information on which sample size estimates could be based.

One goal of many disability measure developers is to come up with a scoring scheme that has interval or ratio measurement properties. Interval and ratio data are desirable because they provide greater information about patients and yield greater statistical power when used as outcome measures in clinical trials. VanSwearingen and Brach have chosen different scoring schemes for the two subscales. Specifically, response option scores vary from 0 to 5 for the physical function subscale and from l to 6 for the social/well-being subscale. Two different transformations are applied to yield scores out of 100. One advantage of using a transformed score is that it provides a method of creating a seemingly meaningful summary score when items have been left blank by the patient. Indeed, this method allows the calculation of a transformed score when only one item has been completed by a patient. Given that the validity of a measure is influenced by its reliability, and one factor affecting reliability is the length of the test, is there a minimum number of items that should be completed before the transformed score can be meaningfully interpreted? A potential difficulty with the physical function subscale scoring scheme is that it assigns a raw score of l to persons who usually do not do an activity because of their health and a raw score of 0 to persons who usually do not do an activity for other reasons. The corresponding transformed score for a person who reports a raw score of l for all five items is 0, compared with--25 for a person who provides a raw score of 0 for all five items. A similar phenomenon also exists for the physical function and social activity scores of the Functional Status Questionnaire.[5] It would be informative to hear in more depth the considerations and methods that have led to this scoring scheme.

As the authors note, one problem facing those evaluating disability measures is that no single agreed-on "gold standard" exists. In such situations, criterion validity gives way to construct validity. In this study, both converging and diverging constructs were formulated and tested.[6](pp155-156) Specifically, the authors hypothesized that the physical function subscale of the FDI should demonstrate a moderate correlation with facial movement evaluated by the impairment-level Facial Grading System (FGS) and that the physical function subscale of the FDI should demonstrate a weak correlation with the physical function subscale of the SF-36. These hypotheses represent examples of convergent and divergent construct validity, respectively. Convergent and divergent validity constructs were also formed to assess the social/well-being subscale. Pearson and Spearman correlation coefficients were used to assess the association between variables identified in the constructs. The Pearson coefficient assumes a linear relationship between the variables under investigation (eg, FGS impairment measure and FDI). Although this method has been used in similar situations, there is growing evidence that the relationship between impairment and disability may not always be linear.[7] It would be interesting to know whether the relationship appears to be linear for facial dysfunction.

The developers of the FDI have designed a measure that is easy to apply and record in the medical record. In clinical practice, the FDI will be used to assess disability at a single point in time and to determine whether change occurs over time. Clinical researchers will likely use the measure to better understand the natural or clinical disability histories of patients with different conditions and as an outcome measure for clinical trials. Moreover, the results of the measure may also prove useful in contributing information when forming an opinion about a patient's prognosis. Each of these potential uses requires a different type of validation study.[8] The current results demonstrate the extent to which face, content, and construct validity exist. As the authors suggest, subsequent inquiry addressing test-retest reliability and sensitivity to clinically important change would seem to be the logical next steps in this exciting and, judging from other measures such as the SF-[36.sup.2] and Roland-Morris Questionnaires,[3,4] never-ending process.

Paul W Stratford, PT Assistant Professor, School of Rehabilitation Science Associate Member, Department of Clinical Epidemiology and Biostatistics McMaster University OT/PT Building T-16 1280 Main St W Hamilton, Ontario, Canada LOS 4KI ([e-mail])

References

[1] Bergner M, Bobbitt RA, Carter WB, Gilson BS. The Sickness Impact Profile: development and final revision of a health status measure. Med Care. 1981;19:787-805. [2] Greenough CG, Fraser RD. Assessment of outcome in patients with low-back pain. Spine. 1992;17:36-41. [3] Roland M, Morris R. A study of the natural history of back pain, part I: development of a reliable and sensitive measure of disability in low-back pain. Spine. 1983;8:141-144. [4] Roland M, Morris R. A study of the natural history of low-back pain, part II: development of guidelines for trials of treatment in primary care. Spine. 1983;8:145-150. [5] Jette AM, Davies AR, Cleary PD, et al. The Functional Status Questionnaire: reliability and validity when used in primary care. J Gen Intern Med. 1986;1:143-149. [6] Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to Their Development and Use. Oxford, England: Oxford University Press; 1995. [7] Buchner DM, de Lateur BJ. The importance of skeletal muscle strength to physical function in older adults. Annals of Behavioral Mediane. 1991;13:91-98. [8] Kirshner N, Guyatt G. A methodological framework for assessing health indices. J Chronic Dis. 1985;38:27-36.

Author Response

In his commentary, Mr Stratford describes the challenges of developing a disability index, including the challenge of further validation of the index. We can attest to Mr Stratford's description of our work as "pioneering." The process of developing and testing a disability index to account for person-level outcomes of what will largely be body system-level interventions in a relatively understudied area of movement disorders was an adventure! As Mr Stratford indicated, previous researchers who have developed or implemented disability indexes (researchers within physical therapy as well as other disciplines) have described concepts, steps, rationales, and some methods of developing and evaluating the validity of a general or a disease-specific disability index. The detailed descriptions and careful referencing included in many of these reports has been extremely helpful in our initial experience with index development. [1-6]

We appreciate Mr Stratford's comments regarding reporting on the Facial Disability Index (FDI). The suggestion to provide the item means and standard deviations for use by subsequent investigators could have been done and would have facilitated the use of the FDI by other clinical investigators, which is a key reason for attempting to publish the study in the public domain. The scoring scheme for transforming scores was taken directly from the scoring method for the Functional Status Questionnaire[5] and was useful because incomplete responses to individual items did not mean discounting the entire patient self-report. On retrospective review of the 46 FDI reports studied, no item received a score of less than 1 and subscale scores were based on responses to all five items of the scale for all except one patient (for that patient, the subscale score was based on four item responses). A strategy, however, for managing zero responses, which could lead to a negative index score (and misinterpretation), and the consideration of a minimum number of items answered for a meaningful subscale score are necessary.

For the initial evaluation of the assessment of disability associated with facial neuromuscular disorders, and without compelling evidence to the contrary, we assumed (as clinicians might assume) that the relationship between impairment and disability would be linear. The preliminary results of recent clinical studies of assessment and outcomes of patients with facial neuromuscular disorders in physical therapy suggest that the relationship may not be truly linear, with various patient factors and conditions modifying the relationship between impairment and disability. In his commentary, Mr Stratford has provided guidance, including highlighting the steps and the types of studies to be undertaken in furthering the validation of the FDI and the clinical usefulness of the disability measure. We appreciate the direction and intend to continue the adventure in the clinical assessment of disability.

Jessie M VanSwearingen, PhD, PT Jennifer S Brach, PT

References

[1] Brook RH, Kamberg CJ. General health status measures and outcome measurement: a commentary on measuring functional status. J Chronic Dis. 1987;40:1315-1365. [2] Deyo RA. Comparative validity of the sickness impact profile and shorter scales for functional assessment in low-back pain. Spine. 1986; 11:951-954. [3] Kirshner B, Guyatt G. A methodological framework for assessing health indices. J Chronic Dis. 1985;38:27-36. [4] Feinstein AR. Clinimetrics. New Haven, Conn: Yale University Press; 1987:35-37, 124-127. [5] Jette AM, Davies AR, Cleary PD, et al. The functional status questionnaire: reliability and validity when used in primary care. J Cen Intern Med. 1986;1:143-149. [6] Waddell G, Main CJ, Morris EW, et al. Chronic low-back pain, psychologic distress, and illness behavior. Spine. 1984;9:209-213.

JM VanSwearingen, PhD, PT, is Assistant Professor, Department of Physical Therapy, School of Health and Rehabilitation Sciences, University of Pittsburgh, 6035 Forbes Tower, Pittsburgh, PA 15260 (USA), and Director of Rehabilitation, Facial Nerve Center, University of Pittsburgh Medical Center. Address all correspondence to Dr VanSwearingen. JS Brach, PT, is Clinical Assistant Professor, Department of Physical Therapy, School of Health and Rehabilitation Sciences, University of Pittsburgh, and Team Leader of Acute Medicine, Department of Physical Therapy, University of Pittsburgh Medical Center.

This study was approved by the Institutional Review Board for Biomedical Research of the University of Pittsburgh.

This article was submitted February 21, 1996, and was accepted August 8, 1996.