Refractive error affects billions of people worldwide, and uncorrected refractive error remains the most common cause of visual impairment and a significant cause of blindness.1–3 Refractive error comprises diverse groups (myopia, hyperopia, astigmatism, or presbyopia) classified based on where the light rays entering the cornea are focused relative to the retina.4 It can be easily corrected by spectacles, contact lenses, or refractive surgery. During the past two decades, refractive surgery has become popular and is one of the most common surgeries performed worldwide.5 Refractive surgery includes corneal (eg, LASIK, photorefractive keratectomy [PRK], laser assisted sub-epithelial keratectomy [LASEK], small incision lenticule extraction [SMILE], and corneal inlays) and intraocular procedures (eg, phakic or pseudophakic intraocular lens [IOL] implantation).6,7
With rapid technological advancements, the outcomes of refractive surgery are now more consistent, stable, and predictable.4,5 However, there are potential complications.4,5,8–12 Visual acuity, contrast sensitivity, residual refractive error, and aberrometry are the objective outcome measures for refractive surgery.13 In addition, patient-reported outcomes are gaining their place as an important part of the comprehensive outcome assessment.14–16 Numerous questionnaires are available for assessing quality of life domains in refractive error,17–31 including those specifically developed for refractive surgery.25–28
The existing questionnaires specific to refractive error are not of equal standard. Moreover, the questionnaires developed to assess outcomes of spectacle wear, contact lens wear, or uncorrected refractive error may not be valid for refractive surgery. Hence, it is difficult to choose an appropriate questionnaire as a clinical or research outcome measure in refractive surgery. Therefore, we performed this systematic review to: (1) identify all of the available questionnaires, (2) determine quality of the existing questionnaires, and (3) evaluate performance of the questionnaires in measuring refractive surgery outcomes.
The relevant articles published before June 19, 2016, were identified by an electronic search in PubMed, MEDLINE, Scopus, CINAHL, Cochrane, and Web of Science databases. The search strategy employed comprehensive search terms (Table 1). It was a systematic and iterative process limiting only to the English language (Figure 1).
Search Keywords for Each Database
Literature review search strategy.
Screening of the articles was done in two stages. First, all abstracts were reviewed. The original articles describing a questionnaire to assess quality of life domain(s) in refractive surgery were included. The studies on cataract surgery were also included if the surgeries were performed for refractive purpose (eg, multifocal IOL implantation).18,31,32 At this stage, the articles were only excluded if they clearly met the exclusion criteria based on the information provided in the abstract (Figure 1). Second, the full text of the potentially relevant articles was retrieved and thoroughly reviewed. The included articles were classified by the types of questionnaires they described. Furthermore, the questionnaires were also grouped as refractive or non-refractive based on the population used for development and validation (Figure 1).33 The non-refractive questionnaires included vision-but-non-refractive questionnaires (questionnaires for general ophthalmic conditions including, but not specific only to, refractive error (eg, the Ocular Surface Disease Index)34 and generic questionnaires (non-disease–specific questionnaires [eg, the McGill Pain Questionnaire]).35
The quality evaluation of the existing questionnaires was performed on content, validity, reliability, responsiveness, and psychometric properties based on both Classical Test Theory (CTT) and Rasch analysis. CTT and Rasch analysis are two commonly used psychometric techniques for developing and validating questionnaires.14,15,36–38 The evaluation was based on an extensive set of quality assessment criteria used previously by our group (Tables A–B, available in the online version of this article).16,39 Each criterion was graded as A (high quality), B (moderate quality), or C (poor quality). These criteria are consistent with the U.S. Food and Drug Administration and the COSMIN (Consensus based Standards for the selection of health status Measurement Instruments) guidelines.40,41
Criteria for grading questionnaires based on the guidelines proposed by Khadka et al1
We identified 81 articles describing 27 questionnaires (12 refractive; 15 non-refractive questionnaires: 7 vision-but-non-refractive, and 8 generic). (Refer to Table C, available in the online version of this article, for a complete list of articles by the types of questionnaire they describe.) Most articles (n = 56, 69.1%) described use of questionnaires specific to refractive error. Others used vision-but-non-refractive (n = 19, 23.5%) and generic (n = 8, 9.9%) questionnaires (Table D, available in the online version of this article). Two articles described more than one type of questionnaire (Figure 1). None of the non-refractive questionnaires were validated in refractive surgery populations.
Non-refractive error-specific questionnaires
We reviewed 56 articles describing development, validation, or application of 12 questionnaires specific to refractive error (Table B). Three questionnaires were developed using Rasch analysis,17–19 and four questionnaires were developed exclusively for refractive surgery.25–28 We extracted information on properties of the questionnaires (frequency of use, concept measured, content development, and scoring), test performance (tests based on CTT or Rasch analysis for assessing psychometric properties, validity, and reliability), and performance of the questionnaire as an outcome measure (eg, responsiveness or sensitivity to the effect of complications). Then the questionnaires were assessed for quality (Table B). If the information on assessment criteria (Table A) is not provided below or in Table B, then either the test was not performed or the information was not provided.
National Eye Institute Refractive Quality of Life (NEI-RQL)
We reviewed 19 articles that describe development or application of the NEI-RQL in various refractive surgery populations, including laser refractive surgery (LASIK, LASEK, or PRK) and toric or astigmatic IOL, multifocal IOL, and phakic IOL implantation. The NEI-RQL has items mainly covering two domains of quality of life (activity limitation and symptoms), although the developers claim it to be a comprehensive measure of quality of life. The NEI-RQL was developed and validated using CTT.20,21 A comprehensive patient consultation was done, but did not report how the final items were selected. Sixteen different response categories are employed for 42 items, making up 13 subscales. The NEI-RQL does not produce an overall score, but a score for each of the 13 subscales. Each subscale has a score from 0 to 100, with higher values indicating favorable outcomes (better quality of life).21
Hays et al.21 reported up to 35% floor effect and up to 82% ceiling effect. The Cronbach α value ranged from 0.64 to 0.90. The test–retest intraclass correlation coefficient (ICC) ranged from 0.55 to 0.83.21 Kobashi et al.42 reported similar results for the Japanese version of the NEI-RQL. The Cronbach α value ranged from 0.61 to 0.90 overall and for all subscales. The ICC ranged from 0.73 to 0.94 overall and for all subscales (Table B).42
McAlinden et al.43 reported that all subscales of the NEI-RQL failed to form valid scales when assessed with Rasch analysis (Table B). Similarly, Blaylock et al.44 found no significant correlation between postoperative refractive error with all NEI-RQL scores. In another study, Iijima et al.45 found moderate correlation (Spearman's correlation coefficient = 0.58) between glare subscale score and scattering from phakic IOL with a central hole (hole ICL).
In general, studies have found better NEI-RQL scores for refractive surgery compared to spectacles or contact lens wear.21,46–49 However, many studies have reported poorer glare subscale scores after refractive surgery.21,48,50,51 The NEI-RQL has been used to compare refractive quality of life in refractive surgical populations. Nichols et al.52 reported significant differences only for 4 of 13 subscale scores between patients with myopia seeking LASIK and patients with myopia not seeking LASIK. McDonnell et al.50 reported similar changes in NEI-RQL scores after keratorefractive surgery for patients with myopia and hyperopia, whereas Blaylock et al.51 reported larger improvement in 2 of 13 subscale scores in patients with hyperopia compared to patients with myopia.
Several studies have compared outcomes of various refractive surgical procedures based on the NEI-RQL subscale scores. Pepose et al.53 found that the bilateral Crystalens group had more favorable outcomes than the combination of Crystalens, ReZoom, or ReSTOR intraocular lenses. Similarly, Kobashi et al.42 found that the phakic IOL group scored better in 4 of 13 subscales than the LASIK group 5 years postoperatively. In another study, the toric IOL implantation group had better NEI-RQL scores for 4 of 13 subscales than the astigmatic IOL implantation group 3 months postoperatively.54 Visser et al.55 and Lin et al.56 observed no significant difference between different IOL implantation combinations. The studies report inconsistent findings for performance of the NEI-RQL subscales in measuring refractive surgery outcomes
Refractive Status and Vision Profile (RSVP)
We reviewed nine articles on the RSVP. The original RSVP is a questionnaire based on the CTT developed almost exclusively for a refractive surgery population. Most of its items are for symptoms and activity limitation domains only. The RSVP produces an overall score and eight subscale scores ranging from 0 to 100. Higher scores indicate more impairment.22,23
The original RSVP was reported to have a good internal consistency (Cronbach α: 0.70 to 0.93). The ICC for the group undergoing refractive surgery and the group not undergoing surgery was 0.61 and 0.88, respectively.22,23 Rasch analysis revealed that the RSVP had underused response categories and poor targeting. The Rasch analysis–guided 20-item (RSVP-20) revised version of the RSVP was then developed, which had an acceptable precision (person separation: 2.01). For the RSVP-20, the Cronbach α value was 0.90 and the ICC was 0.80.57 In another study using Rasch analysis, only two subscales showed adequate measurement precision (> 2.0). However, both subscales suffered from poor targeting and differential item functioning (Table B).58
The Cronbach α value of the Persian RSVP ranged from 0.60 to 0.92. The ICC ranged from 0.51 to 0.95. Most of the RSVP subscale scores weakly correlated with the clinical measures, such as visual acuity and spherical equivalent refractive error. Similarly, the scores for most of the RSVP subscales were not significantly different for those seeking refractive surgery from those who had previously undergone refractive surgery.59
The RSVP has been used as an outcome measure in various types of refractive surgery, including LASIK60 and phakic IOL implantation.61 Schein et al.62 reported that the RSVP is responsive to quality of life changes after refractive surgery (effect size: 1.2 to 1.4). Most patients had significant improvements in overall scale and 2 of 8 subscales. Similarly, Lane and Waycaster61 reported moderate responsiveness of the overall RSVP for phakic IOL implantation for high myopia. The overall effect size value was 0.8 at three postoperative assessments. The effect size for individual subscales ranged from 0.3 to 1.4.61 Similar to the NEI-RQL, the performance of the RSVP in measuring refractive surgery outcomes is inconsistent in the literature.
Quality of Life Impact of Refractive Correction (QIRC)
We identified eight articles on the QIRC. It was developed and validated using CTT and Rasch analysis.17 Comprehensive consultation was done with patients with myopia, hyperopia, and astigmatism, along with literature review and expert opinion. The QIRC has a good coverage of the quality of life domains. Rasch analysis found that patients with refractive correction experience less activity limitation; instead, convenience, health concerns, and emotional and economic issues are more influential on quality of life. With fewer items (n = 20), the QIRC has low respondent burden compared to other widely used refractive questionnaires (NEI-RQL and RSVP). The QIRC score is reported on a converted Rasch scale from 0 to 100. A higher score represents better quality of life, and the average score is close to 50 units.17
The QIRC demonstrated promising psychometric properties such as measurement precision (person separation: 2.03) and fit statistics (infit: 0.70 to 1.24; outfit: 0.78 to 1.32). It is free from floor and ceiling effects. The test–retest ICC was 0.88. Internal consistency (Cronbach α) was 0.78 (Table B).17 For the Greek version of the QIRC, the Cronbach α value ranged from 0.88 to 0.92 for the surgery group, and the ICC was 0.98.63 However, in a recent study, the original QIRC was reported to be multidimensional and it was modified into two unidimensional scales: Functional (items: 1, 3, 7 to 13) and Emotional (items 14, 15, 17 to 19).64
The QIRC has been reported to be responsive to different refractive surgery procedures, including LASIK, LASEK, implantable collamer lens (ICL) implantation, multifocal IOL implantation, and SMILE.17,63–69 In a study by Garamendi et al.,66 the QIRC was responsive to detect change in quality of life after LASIK. Improvement in the scores was observed for all 20 items, of which 16 were statistically significant. Only a small number of patients who had complications had decreased QIRC scores.66 McAlinden and Moore69 and Ieong et al.68 reported improvement in the QIRC score after multifocal IOL implantation and ICL implantation, respectively. Similarly, using the Greek version of the QIRC, Meidani et al.63 found that femtosecond laser–assisted LASIK significantly improves quality of life. Ang et al.64 found no differences between the QIRC scores 3 months after LASIK and SMILE. However, the authors indicated that a study with a larger sample size and a longer follow-up period is required to confirm their findings.64 In another study, Pesudovs et al.65 found that the patients who had refractive surgery had higher QIRC quality of life scores than those who wore spectacles or contact lenses. Likewise, Ieong et al.67 reported higher QIRC quality of life scores for ICL implantation over contact lens wear. To conclude, the QIRC has been proven responsive to changes in refractive surgery outcomes (including complications) and it enables comparison between effectiveness of the newer surgical procedures.
Quality of Vision (QoV)
Five studies were reviewed on the QoV. The content of the QoV was derived from consultation with patients, and its psychometric properties were evaluated using Rasch analysis. It has 30 items for 10 visual symptoms: glare, halos, starbursts, hazy vision, blurred vision, distortion, double or multiple images, fluctuation in vision, focusing difficulties, and difficulty judging distance or depth perception.18 It has three rating scales (Severity, Frequency, and Bothersome). The QoV subscale scores range from 0 to 100. Higher scores indicate poorer quality of vision.18 The three subscales of the QoV are reported to be non-interchangeable.70
Overall, the QoV has excellent psychometric properties. The variance explained by the principal component was greater than 60%. The unexplained variance explained by the first contrast was less than 2.0 eigenvalues for all three scales. Similarly, mean square infit and outfit were within 0.81 to 1.27 and the person separation was greater than 2.0 for all three scales. There was a strong correlation of the QoV scores with visual acuity and contrast sensitivity. The ICC was 0.87. There was differential item functioning for 8 of 30 items. The QoV demonstrated poor targeting.18
Visual symptoms are an important potential complication of refractive surgery.8–12 The QoV has been used to assess symptoms after various types of refractive surgical procedures. McAlinden et al.71 reported worsening of symptoms after LASEK, which subsequently improved to better than the preoperative levels by 3 months postoperatively. Luger et al.72 reported that the QoV scores worsened after presbyopic LASIK before 3 months and remained stable after that. In another study, Maurino et al.73 reported no significant difference between the QoV scores between two types of multifocal IOL implantation (bilateral implantation with the AT LISA 809M IOL or ReSTOR SN6AD1 IOL). A small but clinically significant minority of patients remained symptomatic (particularly halo being more bothersome).73 The QoV performs satisfactorily as an outcome measure in refractive surgery.
Canadian Refractive Surgery Research Group Quality of Vision Questionnaire (QVQ)
We identified four studies in refractive surgery using the QVQ. The QVQ is a CTT-based questionnaire developed to assess patient satisfaction after bilateral PRK surgery.24,25 The items were obtained from the Prospective Evaluation of Radial Keratotomy (PERK) Study questionnaire26 and the Visual Functioning Index (VF-14).26,74 Most of the items are on activity limitations and symptoms. The QVQ employs 5-point Likert scales.25
The Cronbach α value ranged from 0.83 to 0.96. The ICC ranged from 0.21 to 0.92.25
The QVQ was responsive to the PRK,24 LASIK,75 and phakic IOL implantation.76 A high level of satisfaction was reported after each surgery. However, glare and night vision problems were reported to be more problematic.24,75,76
PERK Study Questionnaire
The PERK Study Questionnaire was the first questionnaire (developed in 1986) to evaluate refractive surgery outcomes. It is a CTT-based questionnaire for assessing satisfaction after radial keratotomy. It has items on health concerns, symptoms, and emotional issues.26
The Cronbach α value ranged from 0.89 to 0.90. The item sum correlations were between 0.60 and 0.80.26 Less than half of the participants were satisfied with the radial keratotomy outcomes.26
Multidimensional Quality of Life for Myopia (MQLM) Scale
The MQLM scale is a CTT-based questionnaire developed to assess success of methods for myopia correction. It has items mainly on emotional well-being, symptoms, and activity limitations.29,77 The items were obtained mostly from the literature.29
The Cronbach α value ranged from 0.76 to 0.92. The ICC was 0.75.29
The MQLM scale was responsive to changes in frequency of visual symptoms, psychological state, and overall satisfaction with uncorrected visual acuity after LASIK. However, no statistically significant differences were detected for tolerance of symptoms, cosmesis, and extraversion/introversion subscales.77
Myopia-Specific Quality of Life Questionnaire (MQLQ)
The MQLQ was derived from the preexisting questionnaires to assess quality of life in Korean people who had LASIK for myopia. It is a CTT-based questionnaire that aims to measure quality of life.28 However, it has items mainly on symptoms and activity limitation domains only. All items are rated on a scale ranging from 1 (maximal dysfunction) to 5 (minimal dysfunction).28
The Cronbach α value ranged from 0.70 to 0.95.28 LASIK improved quality of life in myopia. Patients reporting adverse symptoms after LASIK had reduced overall quality of life scores.28
Subjective Vision Questionnaire (SVQ)
The SVQ is a CTT-based questionnaire developed to assess quality of vision in LASIK for myopia. It has items on symptoms and activity limitation domains. It uses a visual analogue scale (10 cm line anchored with descriptive adjectives). Subjective vision index (SVI) is calculated where 0 is very poor and 100 is a perfect SVI.27
The Cronbach α value was 0.94, and the test–retest correlation was 0.79.27 The final 24 items accounted for 67.5% variance on principal component analysis.27 The bilateral Crystalens group had more favorable outcomes than the combination of Crystalens, ReZoom, or ReSTOR IOLs.53
Refractive Error Quality of Life Scale (REQ-Thai)
The REQ-Thai is a CTT-based questionnaire developed to assess refractive-specific quality of life and refractive surgical outcomes in Thai adults.30 The items were obtained from the literature. The original version has 87 items and the new version has 48 items.
The Cronbach α value ranged from 0.74 to 0.99 for five dimensions for the long version and from 0.69 to 0.94 for the short version. The test–retest ICC was 0.92.30
The REQ-Thai has not been used as an outcome measure.
The Freedom from Glasses Value Scale (FGVS)
The FGVS is a CTT-based questionnaire that aims to assess freedom from glasses after multifocal IOL implantation.31,32 It consists of items on health concerns, convenience, and emotional issues. Scores for each item range from 1 to 5, with a higher score meaning a more positive evaluation.
Low missing data reported indicate good acceptability of the FGVS. However, a ceiling effect was observed in all five subscales. The Cronbach α value ranged from 0.78 to 0.93. Scale to scale correlations between five subscales ranged from 0.27 to 0.66, and item to scale correlations ranged from 0.52 to 0.85.31
The participants not wearing glasses had higher scores than those wearing glasses after surgery.31
Near Activity Visual Questionnaire (NAVQ)
The NAVQ was developed to assess outcomes of presbyopic corrections, including various IOLs.19,72 All items are related to activity limitations. The content was developed mostly from the literature.19 The NAVQ has been evaluated using Rasch analysis and CTT. Raw scores are converted to the Rasch scale with higher scores indicating worse visual function.
The NAVQ could discriminate between people with and without near vision difficulty (separation index = 2.92; area under receiver operating characteristic (ROC) curve = 0.91). The correlation coefficient of the questionnaire scores with near visual acuity and critical print size were 0.32 and 0.27, respectively. The ICC was 0.72 and the Cronbach α value was 0.95.19
The NAVQ was responsive to the improvement in the outcomes from the presbyopic LASIK surgery. The scores remained stable after 3 months.72
Quality Assessment of the Refractive Questionnaires
We systematically assessed the quality of the questionnaires (n = 12) validated in refractive surgery (Table B). The QoV18 (grades: 8 As, 2 Bs), the QIRC17 (grades: 7 As, 2 Bs), and the NAVQ19 (grades: 5 As, 3 Bs) are the three best available questionnaires in refractive surgery to assess visual symptoms, quality of life, and activity limitations, respectively (Table B).
We identified 12 refractive error–specific and 15 non-refractive-error–specific questionnaires for assessing quality of life in refractive surgery. None of the non-refractive questionnaires were validated for refractive surgery. Since 1986, numerous refractive questionnaires have been developed. All questionnaires were constructed in developed countries similar to the cataract surgery questionnaires.16 However, unlike for cataract surgery, content for refractive surgery may still be relevant in the developing country settings because refractive surgery is generally accessible only to the people with high socioeconomic status in low income countries.1 However, further research is required to prove or disprove this hypothesis. As with any other questionnaires in optometry and ophthalmology, earlier refractive questionnaires were developed based on CTT.15,16 This includes the most widely used instruments: NEI-RQL20,21 and RSVP.22,23 Use of modern psychometric methods (Rasch analysis) began in refractive surgery with the development of the QIRC in 2004.17
In this review, we comprehensively assessed the quality of all questionnaires based on both CTT and Rasch analysis. Most, including the most frequently used questionnaires (the NEI-RQL and the RSVP), were developed using CTT. Similarly, the questionnaires for measuring quality of life exclusively in refractive surgery (the PERK Study Questionnaire, MQLQ, QVQ, and SVQ) were developed using CTT. The questionnaires developed by CTT provide only the elementary information on psychometrics, validity, and reliability.8,36,37 The NEI-RQL and RSVP were found to be invalid measures when evaluated by Rasch analysis.43,57,58 An attempt to revise the NEI-RQL using Rasch analysis was unsuccessful.43 An attempt to revise the RSVP using Rasch analysis was more successful, although only 2 of 8 subscales were functional and this revised version has not been assessed in outcomes studies.57,58 Despite this, the NEI-RQL and the original RSVP are still being commonly used. The inconsistency in the findings of various studies that have used the RSVP and NEI-RQL probably reflects their poor psychometric properties.
The questionnaires developed by the Rasch analysis (QoV, QIRC, and NAVQ) are of superior quality. This is in agreement with the literature that demonstrates that the questionnaires developed or rescaled using Rasch analysis produce better quality scales than those by CTT.8,36,37,39,63 This is probably because Rasch analysis can identify critical shortcomings and provides opportunity to improve them in a questionnaire. However, even these best questionnaires have limitations. The 20-item QIRC has been reported to be multidimensional, and was split into two unidimensional scales.64 The QoV suffered from poor targeting and differential item functioning.18 The NAVQ had poor item-fit statistics and poor targeting. The dimensionality of the NAVQ has not been reported yet19 (Table B).
The popularity of refractive surgery is growing.4,5 Newer procedures claim to be better. However, quality of life benefits from the newer treatment procedures should be demonstrated.8,16,66 The importance of questionnaires is indisputable for ongoing evaluation of the new technological procedures.8,16,66 From the results of this review, it is evident that the questionnaires may be sensitive to assess the effect of complications on quality of life. However, clinicians and researchers should be careful in choosing a questionnaire. It should be done based on the concept being measured and the quality of the questionnaire, not on the frequency of use or reputation of the developers.37,39 The studies have explored the outcome of refractive surgeries up to 5 years postoperatively. The usefulness of questionnaires in assessing longer-term postoperative outcomes is yet to be studied.
Based on our quality assessment criteria, we recommend the higher quality questionnaires for use. The QoV is the most appropriate questionnaire to assess visual symptoms. The QIRC is the recommended questionnaire to assess quality of life. Similarly, the NAVQ is recommended to measure activity limitation in presbyopia.
The existing questionnaires in refractive error are paper-and-pencil based, and are inflexible. They have a fixed set of items administered to individuals with different levels of ability. Therefore, they either measure low range of trait difficulty or have low precision if they cover a wide range of difficulty levels. They are often poorly targeted to the wide spectrum of refractive error. The shortcomings of the existing questionnaires can be addressed by development of an item bank implemented through a computer adaptive testing system. This technologically advanced dynamic patient-reported outcome measure can offer high measurement precision across wide range of difficulty levels. Using computer adaptive testing, only a few items tailored to individuals are administered thus decreasing respondent burden.15,78 The ‘Eye-tem Bank’ project is currently developing item banks for various ophthalmic conditions including refractive error.15,79