The concept of quality of life encompasses physical, psychosocial, cognitive, and psychological/emotional domains. 1-4 Physical domains include specific symptoms for the disease state being assessed, general vitality, sleep patterns, eating patterns, physical activity, mobility, and self-care. Psychosocial domains include work, community and personal interactions, and intimacy. Cognitive domains include alertness, style of processing information, and problem-solving ability. Psychological and emotional states include anxiety, depression, stress, locus of control, optimism/pessimism, and general sense of well-being. 1-3 Health-related quality of life (HRQOL) is a multidimensional concept that encompasses several areas and not functional status alone. It is influenced by that person's own expectations for well-being, life satisfaction, and happiness. 1,2,4,5 Thus, health-related quality of life is the application of constructs associated with general quality of life to all health conditions. 6
QOL instruments related to visual functioning have focused chiefly on evaluating the effectiveness of treatment options for patients with low vision 7,8 or to assess the quality of life in common sight-threatening conditions such as glaucoma 9 and cataracts. 10,11 The most popular of these are the Visual Functioning Index (VF-14) 12 and versions of the National Eye Institute Visual Function Questionnaire. 13,14 For a full review of these instruments and methods used in questionnaire development, see Massof and Rubin. 15
More common but less severe visual compromise, such as blurred vision or experience with vision correction modalities, can also impact significantly on a person's quality of life. 16-18 Consequently, recent studies have examined HRQOL as it pertains to low, moderate, and high myopia. 16,17 A multidimensional approach to study this problem is justified to help describe, explain, or predict the impact of myopia correction on quality of life. Although new questionnaires have been developed to assess satisfaction and quality of life for persons with refractive errors, 19 their sensitivity in differentiating outcomes among the various treatment options for myopia has been questioned. 20
Psychological Variables in HRQOL
Personality traits 21 and the more transient or situation-dependent psychological states, 22,23 referred to in this article as psychological characteristics, affect overall health and individual response to illness. The Cited Here... describes these terms in more detail. Personality traits such as neuroticism are associated with increased disease susceptibility and the level of symptom reporting. 21 Higher levels of symptom reporting are also associated with extraversion. 21,24-27 Psychological characteristics such as self-efficacy, adaptability, and subjective well-being also influence the ability to cope with physical symptoms. 28 Even rates of recovery are influenced by the degree of self-efficacy, 29 flexibility, 30 and current mood state. 31 Mood states include positive affect and optimism as well as negative affect. 32-34 Research on the effects of optimism on physical well-being supports the belief that optimism is associated with positive mood, coping, and immune functioning. 35 More complaints and susceptibility to stress occur in subjects demonstrating a negative affect. 33
Some theorists maintain that there is a dynamic interaction between personality traits (such as neuroticism) and psychological characteristics such as mood states. 36 However, in a series of studies designed to examine the interaction or independence of psychological characteristics and personality traits, 22 a confirmatory factor analysis yielded a hierarchical factor integrating the psychological characteristics in a health proneness factor.22 Further studies indicated only mild associations between these psychological characteristics and personality traits. 22
The manner in which patients with chronic health conditions, such as cancer, human immunodeficiency virus and acquired immunodeficiency syndrome, heart disease, diabetes, and asthma, experience and cope with the disease and report symptoms is based on psychological characteristics and personality traits. 37-39 Emphasis in the field of health psychology is on the interaction between the physical disease and the psychological variables that affect a person's sense of discomfort in a disease state. 37,39
Psychological Variables and Refractive Correction
Similarly, living with refractive errors and their correction is influenced by individual differences and psychological characteristics, such as adaptability and sense of emotional well-being. 30,40,41 Perceived concern for compromised visual symptoms related to vision correction varies with personality style. 30 Personality traits such as extraversion, introversion, 40 and strength of resolve 41 affect willingness to accept and ability to cope with mild vision compromises. It seems likely that the success of a refractive correction might also be influenced by the person's view of how the correction affects their attractiveness. 42,43 Thus, a scale that integrates personality traits and psychological characteristics with visual compromise should assist practitioners to more broadly understand their patients' needs relative to the correction of myopia. This focus on the multiple dimensions, including psychological factors affecting health, sets this approach apart from others already published 19 or being developed 44,45 in this field.
The aim of this study was to develop a multidimensional HRQOL scale to evaluate perceived problems related to the correction of myopia. In addition to the psychological variables of interest, visually oriented factors such as frequency of visual compromise and ocular symptoms, plus the individual's perception of tolerance for these symptoms, were included. We hypothesized that by examining constructs that can assess individual differences in the tolerance of visual problems within one multidimensional assessment scale, practitioners would be better able to profile their patients' needs, expectations, and personal idiosyncrasies, thereby predicting with better accuracy the success of a given treatment.
PRELIMINARY DEVELOPMENT AND VALIDATION STUDIES
This multidimensional scale development process included development of two independent sets of items (psychologically oriented and visually oriented), modeled on the methods recommended by health-measurement specialists. 46 All stages of the scale development were conducted in accordance with requirements of the relevant ethics review board and with the tenets of the Declaration of Helsinki. Informed consent was obtained from subjects after explanation of the nature of the study. One set of studies focused on examining the health proneness psychological characteristics and personality traits affecting a person's perception of health-related quality of life related to presbyopia and myopia. The second focused on examining specific items in visual compromise and ocular symptoms in myopia correction. Items for the psychological characteristic and personality traits of extraversion and introversion were developed in numerous studies in different sites in both the U.S. and Canada before the final development of the vision items in Australia. Further refinement and factor analysis of the extraversion/introversion items occurred on patient samples in Australia. Integration of the validation process for all psychological items (psychological characteristics and personality traits of extraversion/introversion) and vision items (frequency of visual compromise and ocular symptoms, tolerance of these problems, and cosmesis) occurred after the independent validation of the different groups of items.
Fig. 1 summarizes the pilot studies from which the psychologically and visually oriented items were developed and the procedures used for evidencing validity and reliability for these items. All statistical analyses described were conducted using SPSS for Windows, version 10, 47 except for the confirmatory factor analysis, which was conducted using AMOS. 48
Psychologically Oriented Items
Methods: Content Validity.
A thorough literature review was conducted. Concurrently, focus groups and interviews to elucidate the domains operationally defined in the Cited Here... as psychological characteristics or psychological personality traits were held. These domains included positive and negative affect, optimism, adaptability, flexibility, self-efficacy, self-esteem, extraversion, and introversion. Neuroticism was included under items for negative affect. This led to development of 25 items that examined aspects of these areas.
These 25 items were administered to two graduate education classes (N = 43 and N = 27) for written feedback and preliminary analysis to examine correlations with different psychological measures (e.g., Myers-Briggs Type Inventory 49 and Millon Personality Inventory 50). Items were then administered to subjects attending informational sessions for photorefractive keratectomy (N = 20). The feedback regarding clarity and understanding of the basic meaning of the questions was used to refine the items. The associations between the newly developed items and each of the inventory items were analyzed using a parametric correlation test.
Results: Content Validity.
On the basis of significant correlations across the domains described above, 18 items were selected for further analysis.
Methods: Construct Validity (Exploratory Factor Analysis).
The most common method of evidencing construct validity in scale development is factor analysis. 51 It is considered the most effective method for data reduction and for examining underlying constructs summarizing a set of data. 51 The number of factors representing the underlying constructs should explain ≥50% of the variance. 52 Specific items load on different factors, indicating that the items are correlated with that underlying construct. A commonly acceptable level of factor loading is >0.4, 53 although many consider 0.3 acceptable. 54 The process of deleting items loading below a designated cutoff assists in shortening the scale for the next factor analytical iteration. The process of shortening a scale can continue until the items load solidly on one or another factor above a designated weighting. The weight of the loadings is also important in considering the size of the sample. 53 If the loadings are low (<0.3) and there is a preponderance of items with double loadings (the same variable loading strongly on two factors), then a large sample of ≥300 participants is recommended. However, if after preliminary analyses, the loadings are high (>0.6), a sample of 100 is considered adequate. 55
There are numerous acceptable methods of factor analysis extractions. Maximum likelihood methods are used for reducing a large data set to obtain the maximum likelihood of explaining the variance through the factor analysis. 56 Principal axis factoring is the most common method. 54 Factors' axes are rotated in either orthogonal (perpendicular) or oblique directions to yield a better simple structure that allows for a more clear interpretation of the variables. 57 In the early studies reported here, maximum likelihood was the chosen method when both item and subject numbers were high. Principal axis factoring was used in subsequent studies after the data were reduced.
A convenience sample of 146 subjects between the ages of 41 and 71 years (mean age, 51 ± 6) with presbyopia were recruited at the University of Waterloo optometry clinic. Subjects were participating in a study to examine visual and psychosocial aspects of patient response to different contact lens treatments for the correction of presbyopia. Because presbyopia is an inevitable consequence of aging, the sample was considered representative of the general public and suitable for exploring psychological characteristics and personality traits of people considered psychologically normal.
Exploratory factor analysis was carried out using the 18 items previously selected through examination of content validity. Principal axis factoring was used, 54 and because the factors were hypothesized to be related, oblimin, an oblique method of rotation, was used. 54,56
Results: Construct Validity (Exploratory Factor Analysis).
Eigen values >1 were used to determine the number of factors. 56 Items were retained if factor loadings were >0.3 in the first iteration. A second iteration yielded 10 items on three factors. This analysis explained 62% of the variance. The three factors appeared to explain items that assessed adaptability, self-efficacy, and subjective well-being, which was consistent with the hypothesized construct. Extraversion questions were deleted in this analysis due to low loadings.
Methods: Construct Validity (Confirmatory Factor Analysis).
To further examine the construct structure of the three factors and determine whether there was indeed an overarching construct of health proneness, a confirmatory factor analysis was performed. Subjects were a convenience sample that was seen in a university clinic for routine screening for admission into a range of vision studies. The sample consisted of 341 subjects (145 males and 196 females) who were asked to complete the 18 psychological items described. Confirmatory factor analysis was conducted using the AMOS 48 analysis program and the best-fit model estimated.
Results: Construct Validity (Confirmatory Factor Analysis).
The best-fit model yielded a higher-order health-proneness factor with the original three categories, described above, as subfactors.
Methods: Criterion-Related Validity (Concurrent type).
The Neuroticism, Extraversion, Openness-Five-Factor Inventory, 58 a valid and reliable scale for assessing personality traits was administered to the group involved in the exploratory factor analysis (N = 146) as a criterion validity measure to examine the relationship between the newly developed scale and established constructs of neuroticism, extraversion, openness to experience, conscientiousness, and agreeableness. Associations were examined using a parametric correlation test.
Results: Criterion-Related Validity (Concurrent Type).
The summative total of all new factors was negatively correlated with neuroticism (Pearson r =-0.400, p < 0.01). This indicated that the new scale was assessing a patient's positive and adaptable approach to interactions in the environment, as opposed to hostile and rigid interactions. The summative total was correlated positively with extraversion (Pearson r = 0.437, p < 0.01) and with openness to experience (Pearson r = 0.251, p < 0.01). These results suggested that the newly developed items are only mildly correlated with personality traits, which is consistent with the findings of other studies. 22
Methods: Reliability (Internal Consistency and Test-Retest).
Statistical evaluation of internal consistency of items was performed on the final items yielded by the exploratory factor analysis by examining the Cronbach's alpha coefficient. Test-retest reliability was assessed in 24 subjects (seven males and 17 females; mean age, 33 ± 8 years) who were willing to retake the test items after 6 to 10 weeks. The consistency over time of the total test score was examined using a parametric correlation test. 46
Results: Reliability (Internal Consistency and Test-Retest).
Internal consistency was adequate for the total scale (Cronbach's alpha = 0.78), with individual factors yielding alphas of 0.75, 0.71, and 0.65, respectively. Each individual item was further examined through deletion then rechecking internal consistency. Test-retest reliability was adequate (Pearson r = 0.54). 46
Methods: Construct Validity for Extraversion/Introversion Items.
Extraversion and introversion have been described as both independent of 22 or interdependent with 36 psychological characteristics. To clarify these constructs, six items coding for extraversion/introversion were examined independently of the psychological characteristics.
Two patient samples were selected: a mature age group of 124 subjects (age range, 21 to 56 years; mean age, 38 ± 8) derived from the community and graduate students; the second group was comprised of undergraduate students (N = 118) aged 18 to 26 years. Factor analysis with principle axis factoring was carried out.
Results: Construct Validity for Extraversion/Introversion Items.
Principal axis factoring yielded two constructs (extraversion and introversion) with all items loading >0.4. Strong factors with high loadings suggested that such personality traits are independent constructs and probably not interdependent with psychological characteristics.
Visually Oriented Items
Methods: Content Validity.
In a series of pilot studies to develop items, a thorough literature review of reported vision functioning problems and ocular symptoms related to different treatments for refractive error was carried out. Treatments considered included spectacles, daily wear contact lenses, continuous wear contact lenses, and refractive surgery for myopia. Careful review was made of the National Eye Institute Visual Function Questionnaire test version, 13 the VF-14, 12 and later in the procedure, the National Eye Institute 51-item version 14 and the Refractive Status and Vision Profile 19 to examine the scientific consensus of the items needed to assess the different domains and subdomains within the field of refractive correction. Drafts of items were developed for written feedback on clarity. Periodic revision of the items occurred throughout the validation process.
Small focus groups and personal interviews were conducted with subjects discussing problems and observations with spectacles, contact lenses, and refractive surgery (including photorefractive keratectomy and laser-assisted in-situ keratomileusis [LASIK]). Final items and selection of a representative sample of specific visual compromises and ocular symptoms occurred through these iterative approaches.
Sixty optometry students and academic staff of the Optometry School at the University of New South Wales reviewed the potential items and format of original questionnaires. Written feedback on the specific items and overall format was solicited. Three revisions of the format and items occurred, with the final revisions using a Likert-type scale (e.g., ratings from 1 to 4 with corresponding words for each rating) for each item. Redundancy was examined, and item numbers were reduced accordingly.
Results: Content Validity.
Subjects consistently reported a redundancy of items related to visual function, ocular symptoms, and the corresponding questions related to individual tolerance of these problems. Item numbers were reduced to 37 related to frequency of visual compromise and ocular symptoms and 37 corresponding items examining tolerance levels for the symptom. Six items related to cosmesis were included for a total of 80 questions.
Methods: Construct Validity.
A pilot study was carried out to examine constructs for the vision section of the multidimensional scale. The 80-item scale as described above was administered to 182 myopic subjects between 18 to 40 years of age with refractive errors between -1.00 and -6.00 DS and <1.25 DC. The selection of subjects was chosen to reflect low myopia only. Subjects were recruited from research clinics, optometric practices, and refractive surgery practices. Fifty-seven subjects wore spectacles only (31%), 55 wore daily wear contact lenses (30%), 27 wore continuous-wear contact lenses (15%), 15 had LASIK, and 28 wore a combination of spectacles and daily wear contact lenses (15%). Maximum likelihood exploratory factor analysis with promax rotation was performed, and items with factor loadings <0.4 were deleted.
Results: Construct Validity.
In the preliminary analysis, three iterations yielded a final set of 48 items consisting of 29 items related to the tolerance of ocular symptoms and visual compromise, 14 items related to frequency of ocular symptoms and visual compromise, and five related to cosmesis.
The 48 items derived from the construct validity analysis were subsequently used to compare visual compromise and ocular symptoms between the various treatment modalities in a series of studies described below (Discriminant Validity). Items that were found to be useful in discriminating between treatment groups (i.e., statistically significant and evidencing the utility of the item) are reported in Table 1. In these studies, unadjusted multiple comparisons were used to determine differences between study groups. 59
Methods: Discriminant Validity (Tolerance to Contact Lens Wear).
The scale was administered in a study to examine factors mediating contact lens tolerance in 45 subjects, 24 of whom reported tolerance to daily wear soft lens use (five males and 19 females; mean age, 32 ± 6) and 21 of whom reported intolerance to lens wear (four males and 17 females; mean age, 32 ± 8). Group mean scores for each item were compared using a grouped t-test.
Results: Discriminant Validity (Tolerance to Contact Lens Wear).
Group means were significantly different (p < 0.05) for 23 items, suggesting that specific items were able to discriminate between individuals tolerating and those rejecting contact lenses as a treatment option for myopia (Table 1). Two additional items were marginally significant (p < 0.06).
Methods: Discriminant Validity (Spectacles and Contact Lens Wear).
The scale was administered in a second study to examine the group means between treatment options on 149 myopic subjects with refractive errors between -1.00 and -6.00DS and an age range of 21 to 54 years (mean, 32 ± 7). Subjects were corrected either using spectacles (N = 53; 36%), spectacles and soft contact lenses (N = 27; 17%), daily wear soft contact lenses (N = 60; 40%), or continuous-wear contact lenses (N = 9, 6%). One way analysis of variance was used to assess differences between group means for each of the items.
Results: Discriminant Validity (Spectacles and Contact Lens Wear).
Twenty-two items showed significantly different scores between treatment groups (Table 1).
Methods: Discriminant Validity (Spectacles, Contact Lens Wear, and LASIK).
A post hoc analysis using data collected from 182 subjects reported in the construct validation study was performed. Fifty-seven subjects wore spectacles only (31%), 55 wore daily wear contact lenses (30%), 27 wore continuous-wear contact lenses (15%), 15 had LASIK, and 28 wore a combination of spectacles and daily wear contact lenses (15%). Differences in items between treatment groups were assessed using Wilks' Lambda multivariate test of significance.
Results: Discriminant Validity (Spectacles, Contact Lens Wear, and LASIK).
Nine items were significantly different (p < 0.05) between treatment groups (Table 1). On the basis of these three discriminant validity studies, paired items (frequency of experience of a symptom with the degree of tolerance of the symptom) were retained if they were able to discriminate between treatment groups or between tolerant and intolerant contact lens wearers (Table 1). Additional items, which had factored highly on the loadings and were considered theoretically sound (e.g., difficulty with night vision), were also included.
Of the visual compromise and ocular symptoms items, fifteen matched items (30 items total) with loadings on one factor >0.4 were retained. All six cosmesis questions were retained, although only one item (preferring to be without spectacles in a professional setting) yielded statistical significance in the discriminant analysis. This item selection process yielded 36 questions that were considered useful for the final, integrative factor analytical study.
Methods: Criterion-Related Validity.
Although criterion-related validity is not essential in scale development when there are no instruments assessing exactly what the new one purports to measure, 46 examining relationships between parts of the scale with established measures is useful. 46 A general vision scale (VF-14) 12 and a general physical health scale (Short Form-12) 60 were administered concurrently with the vision items to 62 myopic subjects (mean age, 32 ± 8 years), 24 who had undergone LASIK and 38 who wore daily wear lenses. Associations between the vision items and the VF-14 and Short Form-12 were examined using parametric correlations.
Results: Criterion-Related Validity.
There was a negative correlation (r = -0.528; p < 0.01) between the vision items and VF-14. This association, consistent with the wording of the items, suggested that both scales detected visual compromise. Tolerance of symptoms was negatively correlated (r = -0.311, p < 0.05), indicating that the more subjects were bothered by visual compromise, the more they perceived themselves to have problems on the items assessed by the VF-14. There was a weak association between the cosmesis factor score and the VF-14 (r = 0.26; p < 0.05). There were no significant associations between the vision items and the age- and gender-adjusted subtest scores for mental health and physical health on the Short Form-12.
Methods: Reliability.
Internal consistency was examined using Cronbach's alpha coefficient using the 48-item questionnaire derived from the construct validity analysis. The scale was administered to 182 myopic subjects between 18 and 40 years of age with refractive errors between -1.00 and -6.00 DS and <1.25 DC.
Test-retest correlations were also examined to determine consistency of the scale over time. Twenty-four subjects (seven males and 17 females; mean age, 33 years) were willing to retake the 48-item questionnaire derived from the construct validity analysis. The repeat test was performed 6 to 10 weeks after the initial administration, and a Pearson correlation was used to assess reliability.
Results: Reliability.
Cronbach's alpha coefficients were 0.96 on the tolerance factor, 0.85 on the frequency factor, 0.85 on the cosmesis factor, and 0.91 on the total scale. Each individual item on each factor was further examined through deletion and rechecking of internal consistency. Test-retest reliability was robust (Pearson r = 0.939).
FINAL VALIDATION STUDY (CONSTRUCT VALIDITY)
The aim of this final study was to combine the independently validated psychological items and the visually oriented items into a single scale and to examine construct validity in an unscreened myopic population.
METHODS
Participants
A total of 124 subjects with myopia between -1.00 and -6.00 DS, consisting of 45 men (36%) and 79 women (64%) aged between 21 and 40 years (mean, 31.5 ± 5.7), were seen during routine screening for admittance into a variety of vision studies at The Cooperative Research Center for Eye Research and Technology in Australia. Eighty-eight percent of this sample reported satisfaction with their current form of vision correction. Current correction modalities included contact lenses (N = 73; 59%), a combination of spectacles and contact lenses (N = 9; 7%), or LASIK surgery (N = 42; 34%). Subjects were excluded from the study if they reported symptoms associated with the onset of presbyopia or were over 40 years old. This exclusion criterion was selected to ensure that the final items selected through the factor analysis would be representative of myopic conditions only. This final sample of 124 subjects was selected based on their willingness to answer questions that included psychologically oriented items.
Subjects had a range of 9 to 13 years of schooling with a median duration of 12 years. Seventy-three percent of the patients had some tertiary qualifications either from technical colleges or universities. These qualifications included certificates, trade qualifications, undergraduate training, and postgraduate training such as in dentistry, physiotherapy, and law. Sixty-eight percent worked full time, and 23% worked part time. Although this was a convenience sample, it appears to be representative of people under the age of 40 who live in the eastern suburbs of Sydney, NSW, Australia area as reported in the 1996 census report for population and housing. 61
Procedure
The final version of the multidimensional scale comprising 36 vision-oriented items (15 frequency with 15 corresponding tolerance questions and six cosmesis questions) and 16 psychologically oriented questions (10 psychological characteristics and six extraversion/introversion questions) was administered. Subjects completed the scale after entering the clinic while waiting for the screening examination. They did not know whether they would be eligible for admission to the studies when completing the items for this study. Factor analysis using principal axis factoring was performed to examine construct validity. Oblimin with Kaiser normalization was the rotation method of choice.
Internal consistency for the total scale and subscales was examined using Cronbach's alpha coefficient. Test-retest reliability was determined through intraclass correlation coefficient methods.
RESULTS
In the first iteration, factors were selected using three integrated methods, as previously described. 56 Eigen values (6.77, 4.12, 1.60, 1.18, and 1.11) >1 indicated five factors. The Scree plot flattened after five factors, indicating that no additional factors significantly influenced the model. The five factors identified were consistent with those identified in the preliminary development and validation studies.
The final iteration yielded five factors with 13 items loading on the factor describing tolerance of visual compromise and ocular symptoms, 13 items on the factor describing frequency of symptoms, 10 items on the health proneness (psychological characteristics) factor, six items on the extraversion/introversion (personality traits) factor, and three items on the cosmesis factor. Item loadings ranged from 0.400 to 0.857 (Table 2). Six items below the 0.40 threshold were retained for theoretical purposes until a future study conducted for developing norms based on the proposed scoring method (briefly described in Table 3) is completed. Two matched pairs of visual compromise and ocular symptom items (sore eyes and noticing rim of spectacles) were eliminated due to low item loadings. Three of the cosmesis items were deleted due to low loadings, leaving the minimum acceptable three items. 56 This final factor analysis accounted for 53% of the variance in the data. The final factors were minimally correlated with each other after the final iteration (Table 4).
Cronbach's alpha coefficients were 0.92 for tolerance, 0.82 for psychological states, 0.80 for frequency, 0.80 for extraversion/introversion, 0.76 for cosmesis, and 0.83 for the entire scale, thereby demonstrating strong internal consistency. The intraclass correlation yielded a reliability coefficient of 0.7477.
DISCUSSION
We followed a novel multidimensional approach to describe, explain, and predict the impact of myopia correction on quality of life. We had hypothesized that HRQOL in myopia would be influenced by the frequency and tolerance of visual compromise and ocular symptoms as well as by psychological health proneness characteristics and personality traits. In a series of preliminary studies, items related to each category were independently developed and validated. Finally, the combined scale was tested in a population who were using a range of myopia treatment modalities. Five multidimensional components emerged in a robust scale, describing the hypothesized variables (tolerance of visual and ocular symptoms, health proneness traits, and frequency of visual and ocular symptoms), extraversion/introversion, and cosmesis.
The final scale comprised 13 items measuring the tolerance of visual compromise and ocular symptoms, 13 items measuring the frequency of those symptoms, 10 items measuring health proneness (psychological characteristics), six items measuring extraversion/introversion (personality traits), and three items measuring cosmesis. These factors explained 53% of the variance in the data. Although lower than expected, this proportion of sample variance is considered generally acceptable in health-related scales. 52 The low correlations among factors (Table 4) indicates that these factors were independent constructs.
The combined scale showed good reliability through overall internal consistency, good test-retest reliability for the individual subscales and combined scale, as well as acceptable overall construct validity. This process for determining reliability and validity is supported in other health measurement and psychological scales. 46, 51
A key theoretical objective in selecting items for retention was the maintenance of matched pairs for tolerance and frequency of the symptoms commonly reported by patients using spectacles, wearing contact lenses, or experienced after LASIK surgery. These pairings will eventually yield norms that can assist practitioners not only in assessing the HRQOL in myopia but in selecting treatment options. For example, if a patient has a very low tolerance for a certain problem (e.g., scratchiness) and this occurs frequently in their treatment modality (e.g., contact lenses), there may be other correction methods (e.g., LASIK) that are less frequently associated with that symptom. Another patient with a high tolerance to scratchiness but a low tolerance for blur or variations of vision may be well-suited for contact lenses but reject LASIK.
The item frequency of problems with seeing near detail loaded poorly (0.288), but the corresponding tolerance item loaded high (0.633). It is most likely that this was due to the exclusion of presbyopes in the visually oriented scale development. The near-vision item match was retained for format consistency and to accommodate potential near-vision issues among nonpresbyopes. Further study will be required to assess and, if required, subsequently adapt the scale for use in presbyopes.
Similarly, items for stinging and scratching loaded strongly on the tolerance factor but poorly on the frequency factor. Because the sample for the final factor analysis consisted of persons with generally high satisfaction with their habitual correction, the poor frequency loading is not surprising. Nevertheless, these items discriminated well between individuals who were tolerant and intolerant to contact lenses in the pilot study of discriminant validity. These frequency items were, therefore, retained and were matched with the corresponding tolerance items to maintain format consistency and enhance discrimination between treatment modalities.
Final item selection was also influenced by item performance in the pilot studies. For example, the self-efficacy item yielded a low loading (0.233) on the health proneness factor in the combined construct validity analysis. This item was retained because it factored strongly with this construct in original studies and was an essential variable in the confirmatory factor analysis.
Two items relating to adaptability and to subjective well-being double loaded (weighted strongly) on two factors, health proneness (their intended factor) and extraversion. The items that double loaded described characteristics of adaptability and subjective well-being, respectively. This is consistent with one theory of extraversion, suggesting that individuals scoring highly on extraversion are considered to be highly adaptable as well as self-assured and self-confident. 58 Future studies using the multidimensional scale will examine whether the separate extraversion/introversion factor is essential.
The inclusion of the health proneness items that assess psychological states such as adaptability, self-efficacy, and subjective well-being and the items assessing extraversion will aid the healthcare practitioner in understanding response patterns to treatment options and the manner in which patients communicate. The higher this subscore, the more likely it is that the person will adapt to new methods of treatment intervention 22, 23 and communicate their concerns 21, 24, 30 should any visual or ocular symptoms occur.
The final variable to emerge from the factor analysis was the need to feel attractive in the method of treatment (cosmesis). This factor was found to be a significant discriminator in the preference for contact lenses or surgery over spectacles and has been identified as an important factor in HRQOL of persons with refractive errors. 44, 45
Discriminant validity analysis showed that this scale is able to discriminate between different correction modalities and between tolerant and intolerant contact lens wearers. This sensitivity has not been previously demonstrated for the Refractive Status and Vision Profile, 20 although a recent iteration of the National Eye Institute-Refractive Error Quality of Life has shown both differences between different types of refractive error and between correction modalities. 45 The ability of our scale to discriminate among treatments in a low (up to -6.00 D) myopic population suggests that it should readily discriminate among treatments in high myopia where visual and ocular symptoms are more frequently reported and more bothersome. 17
A key strength of this new scale is its multidimensionality. No vision scales published to date (i.e., Refractive Status and Vision Profile 19 and National Eye Institute-51 14) or being developed (e.g., National Eye Institute-Refractive Error Quality of Life 45) have addressed the psychological variables known to influence a patient's health-related quality of life. Moreover, the psychological items included in the scale are specifically designed to be administered by a trained healthcare practitioner, requiring a psychologist for neither administration nor interpretation.
Future studies will examine the relationship between an individual subject's tolerance of symptoms and the effect this has on the level of satisfaction with treatment outcomes. Sensitivity and specificity of each item will be specifically examined in the process of developing normative standards for this multidimensional quality-of-life scale for myopia.
In summary, the Institute for Eye Research multidimensional quality of life for myopia scale assesses five dimensions that are relevant to choosing treatment options for myopia. Any trained employee can rapidly and easily administer it in the practitioners' office. The items, scoring system, and interpretation are available from the authors on request. The combination of visually oriented items with health proneness, cosmesis, and extraversion items allows practitioners a method to understand the needs of their patients in terms of the frequency of their symptoms, their tolerance for those symptoms, the importance of a cosmetically desirable option, their overall sense of psychological well-being, and the manner in which they communicate their needs. Ultimately, we expect this approach to prove valuable in providing enhanced success and patient satisfaction by aiding in the selection of the most appropriate treatment.
ACKNOWLEDGMENTS
Supported, in part, by the Australian Government through the Cooperative Research Centres Programme and by the Institute for Eye Research.
We thank Associate Professor Deborah Sweeney and Professor Des Fonn for their contributions.
APPENDIX
Psychological Characteristics
Psychological characteristics are state-dependent behaviors that are believed to contribute to the manner in which persons adjust to and cope with different situations. 62 For example, subjective well-being encompasses the sense of feeling positive or negative or optimistic or pessimistic about a situation being experienced. Self-efficacy is the degree to which a person believes that he/she will succeed in a given situation or treatment. Adaptability is the ability to manipulate or adjust to manage in a new environment or physical condition, often through development of alternative problem-solving strategies. These characteristics have been found to be relevant in health (health proneness). 22
Psychological Personality Traits
Psychological personality traits are theoretical constructs that are believed to be embedded in a person's behavior, consistent across situations and over time. Neuroticism and extraversion/introversion, commonly considered to fit this pattern, have been found to affect how persons perceive and report symptoms to healthcare professionals. 21, 25, 27, 58
Psychological State vs. Psychological Trait
Competing and conflicting psychological theories abound regarding the attribution of patients' behaviors in health conditions as personality embedded or situation dependent. Recent health psychology literature suggest a trend to accept that individual patient response patterns are influenced by a combination of traits and states (e.g., neuroticism, mood states, and situational states). 36, 62, 63
Scale Development
Sources of Scale Items.
Included in the development of items is the qualitative process of gathering information through focused discussions among a sample of people from the target group, loosely guided by a facilitator to elicit spontaneous issues related to the target condition. Specific groups or individual discussions involve key informants (e.g., clinicians and patients) who can lend expert opinion from extensive knowledge of the condition being assessed. In addition, the clinical observations from the researcher's own experience and a thorough literature review on other evidenced-based observations are considered in the final development of specific items. 46
Content Validity.
Establishing content validity assures that the items assess domains of the intended field of study. This is accomplished through discussions with expert panels to seek consensus on the domains to be measured and focus groups and literature reviews to guide item development. Only items related to the content domains are retained. 46
Construct Validity.
Construct validity assures the theoretical integrity of the domains to be measured. Construct validity is commonly established using factor analysis to statistically confirm coherence of the theoretical entities (the factors) involved in the development of the scale. 51 Construct validity may also include assessment of how well the scale relates to other assessment formats (convergent validity) and how effectively the scale discriminates among conditions or treatments (discriminant or divergent validity). Wilks' Lambda is a common statistical method used for establishing statistical significance between conditions or treatments. 46, 57
Criterion Validity.
Criterion validity is assessed by comparing the assessment of persons using the new scale against some generally accepted measure (gold standard) of related concepts that are important in the condition of interest. Criterion validity may also include assessment of how well the scale predicts outcomes for the condition of interest (predictive validity) or a comparison of simultaneous performance on the new scale and a selected standard (concurrent validity). The existence of strong criterion validity, however, suggests that the new scale provides much the same information as the standard and is, therefore, considered to have limited value in the development of health-related scales. 46
Reliability.
Reliability is the consistency of an assessment tool over time. Common statistical analyses applied to scale reliability include Cronbach's alpha correlations for internal consistency and Pearson product-moment correlation for test-retest scores. Cronbach's alpha correlations should be ≥0.8 for internal consistency and ≥0.5 for test-retest. 46 Test-retest correlations must, however, be treated with considerable caution, 46 especially in healthcare settings. If the test-retest interval is short, there is a high risk that patient memory of the test will influence responses. If the test-retest interval is long, there is a high risk that the progress of the condition or the effects of treatment will influence responses. Cited Here...
REFERENCES
1. Bergner M. Quality of life, health status, and clinical research. Med Care 1989; 27: S148-56.
2. Faden R, Leplege A. Assessing quality of life: moral implications for clinical practice. Med Care 1992; 30: S166-75.
3. Muldoon MF, Barger SD, Flory JD, Manuck SB. What are quality of life measurements measuring? BMJ 1998; 316: 542-5.
4. Gold M, Franks P, Erickson P. Assessing the health of the nation: the predictive validity of a preference-based measure and self-rated health. Med Care 1996; 34: 163-77.
5. Hoffman LG, Rouse MW, Brin BN. Quality of life: a review. J Am Optom Assoc 1995; 66: 281-9.
6. Gill TM, Feinstein AR. A critical appraisal of the quality of quality-of-life measurements. JAMA 1994; 272: 619-26.
7. Brenner MH, Curbow B, Javitt JC, Legro MW, Sommer A. Vision change and quality of life in the elderly: response to cataract surgery and treatment of other chronic ocular conditions. Arch Ophthalmol 1993; 111: 680-5.
8. Ellwein LB, Fletcher A, Negrel AD, Thulasiraj RD. Quality of life assessment in blindness prevention interventions. Int Ophthalmol 1994; 18: 263-8.
9. Parrish RK Jr. Visual impairment, visual functioning, and quality of life assessments in patients with glaucoma. Trans Am Ophthalmol Soc 1996; 94: 919-1028.
10. Legro MW. Quality of life and cataracts: a review of patient-centered studies of cataract surgery outcomes. Ophthalmic Surg 1991; 22: 431-43.
11. Javitt JC, Wang F, Trentacost DJ, Rowe M, Tarantino N. Outcomes of cataract extraction with multifocal intraocular lens implantation: functional status and quality of life. Ophthalmology 1997; 104: 589-99.
12. Steinberg EP, Tielsch JM, Schein OD, Javitt JC, Sharkey P, Cassard SD, Legro MW, Diener-West M, Bass EB, Damiano AM, Steinwachs DM, Sommer A. The VF-14: an index of functional impairment in patients with cataract. Arch Ophthalmol 1994; 112: 630-8.
13. NEI-VFQ Phase 1 Development Team. Measuring Visual Functioning: Test Version of the NEI-VFQ. Santa Monica, CA: RAND, 1995.
14. Mangione CM, Berry S, Spritzer K, Janz NK, Klein R, Owsley C, Lee PP. Identifying the content area for the 51-item National Eye Institute Visual Function Questionnaire: results from focus groups with visually impaired persons. Arch Ophthalmol 1998; 116: 227-33.
15. Massof RW, Rubin GS. Visual function assessment questionnaires. Surv Ophthalmol 2001; 45: 531-48.
16. Brown GC. Vision and quality-of-life. Trans Am Ophthalmol Soc 1999; 97: 473-511.
17. Rose K, Harper R, Tromans C, Waterman C, Goldberg D, Haggerty C, Tullo A. Quality of life in myopia. Br J Ophthalmol 2000; 84: 1031-4.
18. Wensor M, McCarty CA, Taylor HR. Prevalence and risk factors of myopia in Victoria, Australia. Arch Ophthalmol 1999; 117: 658-63.
19. Vitale S, Schein OD, Meinert CL, Steinberg EP. The refractive status and vision profile: a questionnaire to measure vision-related quality of life in persons with refractive error. Ophthalmology 2000; 107: 1529-39.
20. Nichols JJ, Mitchell GL, Zadnik K. The performance of the refractive status and vision profile survey in a contact lens clinical trial. Ophthalmology 2001; 108: 1160-6.
21. Feldman PJ, Cohen S, Doyle WJ, Skoner DP, Gwaltney JM Jr. The impact of personality on the reporting of unfounded symptoms and illness. J Pers Soc Psychol 1999; 77: 370-8.
22. Bernard LC, Hutchison S, Lavin A, Pennington P. Ego-strength, hardiness, self-esteem, self-efficacy, optimism and maladjustment: health related personality constructs and the Big Five model of personality. Assessment 1996; 3: 115-31.
23. Friedman HS, Booth-Kewley S. The disease-prone personality: a meta-analytic view of the construct. Am Psychol 1987; 42: 539-55.
24. Erickson DB, Stapleton F, Erickson PM, Kleitman S, Cengiz A, Glasson M, Skotnitsky C, Holden B. Clinical variables and personality factors mediate contact lens tolerance. Invest Ophthalmol Vis Sci 2000; 41: S927.
25. Wade JB, Dougherty LM, Hart RP, Rafii A, Price DD. A canonical correlation analysis of the influence of neuroticism and extraversion on chronic pain, suffering, and pain behavior. Pain 1992; 51: 67-73.
26. Hahn SR, Thompson KS, Budner NS, Stern V, Wills TA. The difficult doctor-patient relationship: somatization, personality and psychopathology. J Clin Epidemiol 1994; 47: 647-57.
27. Harkins SW, Price DD, Braith J. Effects of extraversion and neuroticism on experimental pain, clinical pain, and illness behavior. Pain 1989; 36: 209-18.
28. Barsky AJ, Wyshak G. Hypochondriasis and somatosensory amplification. Br J Psychiatry 1990; 157: 404-9.
29. Schwarzer R, Fuchs R. Self-efficacy and health behaviors. In: Conner M, Norman P, eds. Predicting Health Behavior: Research and Practice with Social Cognition Models. Buckingham, UK: Open University Press, 1996; 163-96.
30. Erickson DB, Ryan RA, Erickson P, Aquavella JV. Cognitive styles and personality characteristics strongly influence the decision to have photorefractive keratectomy. J Refract Surg 1995; 11: 267-74; discussion 274-81.
31. DeNeve KM, Cooper H. The happy personality: a meta-analysis of 137 personality traits and subjective well-being. Psychol Bull 1998; 124: 197-229.
32. Cohen S, Doyle WJ, Skoner DP, Fireman P, Gwaltney JM Jr, Newsom JT. State and trait negative affect as predictors of objective and subjective symptoms of respiratory viral infections. J Pers Soc Psychol 1995; 68: 159-69.
33. Kvaal SA, Patodia S. Relations among positive affect, negative affect, and somatic symptoms in a medically ill patient sample. Psychol Rep 2000; 87: 227-33.
34. Kressin NR, Spiro A III, Skinner KM. Negative affectivity and health-related quality of life. Med Care 2000; 38: 858-67.
35. Segerstrom SC, Taylor SE, Kemeny ME, Fahey JL. Optimism is associated with mood, coping, and immune change in response to stress. J Pers Soc Psychol 1998; 74: 1646-55.
36. Larsen RJ, Ketelaar T. Personality and susceptibility to positive and negative emotional states. J Pers Soc Psychol 1991; 61: 132-40.
37. Yamaoka K, Shigehisa T, Ogoshi K, Haruyama K, Watanabe M, Hayashi F, Hayashi C. Health-related quality of life varies with personality types: a comparison among cancer patients, non-cancer patients and healthy individuals in a Japanese population. Qual Life Res 1998; 7: 535-44.
38. Schneiderman N, Antoni MH, Saab PG, Ironson G. Health psychology: psychosocial and biobehavioral aspects of chronic disease management. Annu Rev Psychol 2001; 52: 555-80.
39. Baum A, Posluszny DM. Health psychology: mapping biobehavioral contributions to health and illness. Annu Rev Psychol 1999; 50: 137-63.
40. Erickson DB, Erickson P. Psychological factors and sex differences in acceptance of monovision. Percept Mot Skills 2000; 91: 1113-9.
41. du Toit R, Ferreira JT, Nel ZJ. Visual and nonvisual variables implicated in monovision wear. Optom Vis Sci 1998; 75: 119-25.
42. Terry RL. The psychology of visual correctives: social and psychologic effects of eyeglasses and contact lenses. Optom Mon 1982; 73: 137-42.
43. Terry RL. Eyeglasses and gender stereotypes. Optom Vis Sci 1989; 66: 694-7.
44. Pesudovs K, Elliott DB, Gelsthorpe SD, Garamendi E, Burr JM. The development of the quality of life impact of refractive correction questionnaire. Invest Ophthalmol Vis Sci 2002; 43: S3814.
45. Hays RD, Mangione CM, Ellwein L, Lindblad AS, Spritzer KL, McDonnell PJ. Psychometric properties of the National Eye Institute Refractive Error Quality of Life (NEI-RQL) instrument. Ophthalmology 2003; 110: 2292-301.
46. Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to Their Development and Use, 2nd ed. London: Oxford University Press, 1995.
47. Norusis MJ. SPSS for Windows Base System User's Guide. Chicago: SPSS, 1996.
48. Arbuckle JL. AMOS: Version 3.6 (w32). Chicago: SmallWaters, 1994.
49. Myers I, McCaulley MH. Manual: A Guide to the Development and Use of the Myers-Briggs Type Indicator (MBTI). Palo Alto, CA: Consulting Psychologists Press, 1992.
50. Millon T. Millon Index of Personality Types (MIPS) Manual. San Antonio, TX: Psychological Corporation, 1994.
51. Clark LA, Watson D. Constructing validity: basic issues in objective scale development. Psychol Assess 1995; 7: 309-19.
52. Streiner DL. Figuring out factors: the use and misuse of factor analysis. Can J Psychiatry 1994; 39: 135-40.
53. Floyd FJ, Widaman KF. Factor analysis in the development and refinement of clinical assessment instruments. Psychol Assess 1995; 7: 286-99.
54. Tabachnick BG, Fidell LS. Using Multivariate Statistics, 3rd ed. New York: Harper Collins College Publishers, 1996: 635-706.
55. Guadagnoli E, Velicer WF. Relation of sample size to the stability of component patterns. Psychol Bull 1988; 103: 265-75.
56. Kim JO, Mueller CW. Factor Analysis: Statistical Methods and Practical Issues. Newbury Park, CA: Sage Publications, 1978: 1-88.
57. George D, Maller P. SPSS for Windows: Step by Step, 4th ed. Boston: Allyn & Bacon, 2003.
58. Costa PT, McCrae RR. NEO PIR/FFI Professional Manual. Odessa, FL: Psychological Assessment Resources, 1992.
59. Cook RJ, Farewell VT. Multiplicity considerations in the design and analysis of clinical trials. J R Stat Soc 1996; 159: 93-110.
60. Andrews G. A brief integer scorer for the SF-12: validity of the brief scorer in Australian community and clinic settings. Aust NZ J Public Health 2002; 26: 508-10.
61. Australian Bureau of Statistics. Census Data 1996. Available at:
http://www.abs.gov.au. Accessed July 2001.
62. Diener E. Traits can be powerful but are not enough: lessons from subjective well-being. J Res Personality 1996; 30: 389-99.
63. Mischel W. On the interface of cognition and personality: beyond the person-situation debate. Am Psychologist 1979; 34: 740-54.