Old measures and new scores in spinal muscular atrophy patients

Introduction: A recent Rasch analysis performed on the Hammersmith Functional Motor Scale—Expanded (HFMSE) in patients with spinal muscular atrophy (SMA) identified issues impacting scale validity, redundant items, and disordered thresholds on some items. Methods: We modified the HMFSE scoring based on the Rasch analysis and on expert consensus to establish whether the traditional scoring overestimated the number of patients with changes within 2 points from baseline. Data were collected retrospectively from multicenter data sets in 255 type 2 and 3 SMA patients. Results: The mean 12‐month changes using the new and the traditional scoring system did not differ significantly (P > 0.05). The numbers of patients who improved or decreased by >2 points were also similar. Conclusions: The presence of outliers using the traditional scoring system was not due to overestimation of changes in activities that were tested bilaterally or to discrepancies in the scoring hierarchy of individual items. Muscle Nerve 52:435–437, 2015     

A new look at the construct validity of the K6 using Rasch analysis

The Kessler six‐item psychological distress (K6) scale is widely used to screen for mental disorders; however, information is lacking on the rating scale performance or dimensionality structure of the scale. This study used a population based sample (n = 7596) to evaluate the construct validity of the K6 scale using Rasch partial credit analysis. The analysis showed that almost all of the five‐point rating scales in the K6 items were used appropriately to differentiate psychological distress of the study participants. The analysis provided evidence of unidimensionality of the scale, although items 1 (so sad) and 3 (restless or fidgety) might offer a potential second off‐dimensional component. All items appeared to fit the Rasch model's expectation as demonstrated by the acceptable item fit statistics. The study participants demonstrated valid response patterns when answering K6 items, except for some who were younger or had higher psychological distress. This study using Rasch analysis confirms the construct validity of the K6 scale and suggests that the K6 is a useful and valid instrument for assessing psychological distress in the mid‐aged general population. Further research can facilitate better understanding about the unidimensionality of the scale. Copyright © 2014 John Wiley & Sons, Ltd.     

Rasch analysis of the Pediatric Evaluation of Disability Inventory–computer adaptive test (PEDI‐CAT) item bank for children and young adults with spinal muscular atrophy

Introduction: In this study we evaluated the suitability of a caregiver‐reported functional measure, the Pediatric Evaluation of Disability Inventory–Computer Adaptive Test (PEDI‐CAT), for children and young adults with spinal muscular atrophy (SMA). Methods: PEDI‐CAT Mobility and Daily Activities domain item banks were administered to 58 caregivers of children and young adults with SMA. Rasch analysis was used to evaluate test properties across SMA types. Results: Unidimensional content for each domain was confirmed. The PEDI‐CAT was most informative for type III SMA, with ability levels distributed close to 0.0 logits in both domains. It was less informative for types I and II SMA, especially for mobility skills. Item and person abilities were not distributed evenly across all types. Conclusions: The PEDI‐CAT may be used to measure functional performance in SMA, but additional items are needed to identify small changes in function and best represent the abilities of all types of SMA. Muscle Nerve 54 : 1097–1107, 2016     

Fatigue and daytime sleepiness scale in myotonic dystrophy type 1

Introduction: Fatigue and excessive daytime sleepiness are frequent complaints in myotonic dystrophy type 1 (DM1) that often overlap. We aimed to construct a combined fatigue and daytime sleepiness rating scale for DM1 using the Rasch measurement model. Methods: Questionnaires, including the Epworth Sleepiness Scale, Fatigue Severity Scale, and Daytime Sleepiness Scale, were completed by 354 patients. Data were subjected to Rasch analyses and tested for required measurement issues such as appropriate response categories, absence of item bias, local independence, and unidimensionality. Results: The initial 22 items did not meet Rasch model expectations. After rescoring and removing misfitting items, the final 12‐item scale showed good model fit and unidimensionality. High internal consistency (person separation index = 0.80) and validity were demonstrated. Conclusions: The Rasch‐built Fatigue and Daytime Sleepiness Scale, developed specifically for DM1 patients, provides interval measures on a single continuum. Its use is suggested for future clinical trials and therapeutic follow‐up. Muscle Nerve, 2013     

International clinimetric evaluation of the MG‐QOL15, resulting in slight revision and subsequent validation of the MG‐QOL15r

Introduction: The MG‐QOL15 is a validated, health‐related quality of life (HRQOL) measure for myasthenia gravis (MG). Widespread use of the scale gave us the opportunity to further analyze its clinimetric properties. Methods: We first performed Rasch analysis on >1,300 15‐item Myasthenia Gravis Quality of Life scale (MG‐QOL15) completed surveys. Results were discussed during a conference call with specialists and biostatisticians. We decided to revise 3 items and prospectively evaluate the revised scale (MG‐QOL15r) using either 3, 4, or 5 responses. Rasch analysis was then performed on >1,300 MG‐QOL15r scales. Results: The MGQOL15r performed slightly better than the MG‐QOL15. The 3‐response option MG‐QOL15r demonstrated better clinimetric properties than the 4‐ or 5‐option scales. Relative distributions of item and person location estimates showed good coverage of disease severity. Conclusions: The MG‐QOL15r is now the preferred HRQOL instrument for MG because of improved clinimetrics and ease of use. This revision does not negate previous studies or interpretations of results using the MG‐QOL15. Muscle Nerve 54 : 1015–1022, 2016     

The Washington Psychosocial Seizure Inventory (WPSI): psychometric evaluation and future applications.

The Washington Psychosocial Seizure Inventory (WPSI) clinical scales were developed via an empirical item selection approach and have been used widely to measure aspects of psychosocial functioning of patients with epilepsy. However, these empirically derived clinical scales have not been assessed psychometrically using a modern item response theory-based model. The goals of this study were to: (1) evaluate how items in each clinical scale performed in such a way as to represent the underlying constructs being measured; and (2) derive a shorter version while maintaining measurement precision. WPSI item response data from 145 adults with epilepsy collected for an evaluation study of an intervention to pact negative attributional style in epilepsy were used. The dichotomous Rasch model suitable for the true-false response choices was used to analyse each clinical scale separately. Most items within each scale fit the measurement model well, with very few exceptions. All items, therefore, were retained. A method, based on computerised adaptive testing (CAT), is offered for shortening the WPSI using a psychosocial outcomes item bank derived from the study. Individuals' specific levels of functioning are used to derive measures of their psychosocial functioning with a minimum number of items.     

The construct validity of the Perceived Stress Scale

ObjectiveStress impacts the quality of life and is associated with increased risk of mental and physical disorders. The Perceived Stress Scale (PSS) is widely used for measuring psychological distress. Although the instrument was originally defined as a single construct, several studies based on classical test theory suggest that a two-dimensional structure is more dominant. We aimed to explore the construct validity and dimensionality of the PSS-10 using modern test theory to determine if the scale is predominantly for a one- or a two-dimensional model.MethodsThe study population consisted of 32,374 citizens who completed the PSS-10 as part of the Danish National Health Survey in 2010. We investigated the construct validity of the PSS-10 by CFA. We examined the scalability by investigating the fit of the data distribution in a unidimensional Rasch model and performing modification of response categories, persons and items. The scale dimensionality was additionally assessed by Mokken and Rasch analysis.ResultsThe PSS-10 did not fit the Rasch model. Item four indicated the largest misfit, and items four and seven displayed disordered thresholds. Unidimensionality could not be established although the data showed improved fit to the Rasch model for the two dimensions respectively with the positive and negative items. Mokken analysis revealed fit to the unidimensional model, but disordered thresholds were shown for item four.ConclusionOur large population-based study indicated scalability problems in the current version of the PSS-10. The conducted analysis overall revealed better statistical fit for a two-dimensional than a unidimensional model.     

Moving towards meaningful measurement: Rasch analysis of the North Star Ambulatory Assessment in Duchenne muscular dystrophy

Aim Reliable measurement of disease progression and the effect of therapeutic interventions in Duchenne muscular dystrophy (DMD) require clinically meaningful and scientifically sound rating scales. Therefore, we need robust evidence to support such tools. The North Star Ambulatory Assessment (NSAA) is a promising, clinician‐rated scale with potential uses spanning clinical practice and clinical trials. In this study, we used Rasch analysis to test its suitability in these roles as a measurement instrument. Method NSAA data from 191 ambulant boys (mean age at assessment 7y 8mo, SD 2y 4mo; range 3y 6mo–15y 5mo) with a confirmed diagnosis of DMD were examined for psychometric properties including clinical meaning, targeting, response categories, model fit, reliability, dependency, stability, and raw to interval‐level measurement. All analyses were performed using the Rasch Unidimensional Measurement Model. Results Overall, Rasch analysis supported the NSAA as being a reliable (high Person Separation Index of 0.91) and valid (good targeting, little misfit, no reversed thresholds) measure of ambulatory function in DMD. One item displayed misfit (lifts head, fit residual 6.9) and there was evidence for some local dependency (stand on right/left leg, climb and descend box step right/left leg, and hop on right/left leg, residual correlations >0.40), which we provide potential solutions for in future use of the NSAA. Importantly, our findings supported good clinical validity in that the hierarchy of items within the scale produced by the analyses was supported by clinical opinion, thus increasing the clinical interpretability of scale scores. Interpretation In general, Rasch analysis supported the NSAA as a psychometrically robust scale for use in DMD clinical research and trials. This study also demonstrates how Rasch analysis is a useful instrument to detect and understand the key measurement issues of rating scales.     

Painful Diabetic Neuropathy Anxiety Rasch‐Transformed Questionnaire (PART‐Q30©)

The association between painful diabetic neuropathy (PDN) and anxiety has been acknowledged using various anxiety scales capturing various fear entities. It has never been examined whether these generally applied anxiety questionnaires could be pooled to construct one overall anxiety metric. After completion by a cohort of 151 patients with PDN, data obtained from seven generally applied fear scales were stacked (n = 88 items) and subjected to Rasch analyses (pre‐PART‐Q88) to create the PDN overall Anxiety Questionnaire (PART‐Q30^©). We subsequently examined the impact of the final constructed PART‐Q30^© on disability and Quality of Life (QoL) using the Rasch‐Transformed Pain Disability Index (RT‐PDI) and the Norfolk Quality of Life Questionnaire, Diabetic Neuropathy version (RT‐Norfolk). The pre‐PART‐Q88 data did not meet Rasch model's expectations. Through stepwise examination for model fit, disordered thresholds, local dependency and item bias, we succeeded in reducing the data and constructing a 30 items overall anxiety scale (PART‐Q30^©) that fulfilled all model's expectations, including unidimensionality. An acceptable internal reliability was found (person‐separation‐index: 0.90). PART‐Q30^© explained 36% of disability and combined with RT‐PDI 63% of QoL (assessed with RT‐Norfolk).     

The SCL-90 and SCL-90R versions validated by item response models in a Danish community sample

OBJECTIVE: To evaluate the internal validity of the subscales of the combined SCL-90 and SCL-90R, the SCL-92, by item response analyses as compared with several previously reported factor analyses of this questionnaire in the literature. METHOD: The SCL-92 questionnaire was mailed to an age- and gender-stratified random sample of Danish citizens. The sample comprised 2040 individuals. The internal structure of the nine factors of the SCL-92 questionnaire was evaluated by Mokken-Loevinger analysis and Rasch analysis. RESULTS: In total, 1153 persons or 58% returned the questionnaire fully completed. Mokken analysis found all scales apart from the psychoticism scale acceptable. The Rasch analysis found most of the subscales to be robust. Minor problems were seen for the scales of phobic anxiety, obsession-compulsion and depression. Analysis of the Global Severity Index showed that the Rasch model was rejected for the full 92-item scale, but not for a scale consisting of the 63 items from the non-psychotic subscales. Spearman correlations among the subscales were all positive (range 0.34-0.79) and so were correlations between each of the subscales and the Global Severity Index (range 0.55-0.91). CONCLUSION: In this sample from the Danish general population the non-psychotic subscales, i.e. the subscales covering psychological distress were observed to function well. In a general population sample, the 63 non-psychotic items primarily appear to reflect one broad dimension of distress.     

Evaluation of the structure of Brazilian State-Trait Anxiety Inventory using a Rasch psychometric approach

Objective

This study evaluates the State-Trait Anxiety Inventory (STAI) structure using a Rasch psychometric approach, and a refined and shorter STAI version is proposed.

Methods

A cross-sectional study was performed with 900 inpatients scheduled for elective surgery. Age varied from 18 to 60 years (American Society of Anesthesiologists physical status I-III). Demographic information was collected using a structured questionnaire. The measuring instrument (the STAI) was applied to all patients in the afternoon before the surgery and prior to the patients receiving preoperative sedatives.

Results

Rasch analysis of the state and trait anxiety scales was performed separately. This analysis demonstrated that the original format of state and trait scales fails to show invariance across the trait-state anxiety level, which results in the unstable performance of items. The refined scale was retested in two subsequent random samples of 300 subjects each, and the results were confirmed. The performance was adequate regardless of gender. In the analysis, some items of the state scale (items 3,4,9,10,12,15, and 20) were deleted due to poor fit statistics. The remaining 13 items showed unidimensionality, local independence, and adequate index of internal consistency. Also, the original trait scale displayed several weaknesses. First, the four-point Likert response scale proved to be inadequate, and threshold disorders were found in all 20 items. Also, the original trait scale showed insufficient item-trait interaction and several individual item misfits. Following the rescoring process, and retesting in a second random sample, items were excluded (namely Items 3, 4, 11, 13, 14, 15, 18, and 19). The refined version showed local independence, unidimensionality, and adequate fit statistics.

Discussion

The results indicate that the application of the Rasch model led to the refinement of the classic STAI state and trait scales. In addition, they suggest that these shorter versions have a more suitable psychometric performance and are free of threshold disorders and differential item functioning problems.     

Psychometric evaluation of the myasthenia gravis composite using Rasch analysis

Introduction: The MG Composite (MGC) is a validated outcome measure of the clinical manifestations of myasthenia gravis. Methods: We performed Rasch analyses of the MGC to investigate additional properties, including its unidimensionality and the appropriateness of the weights assigned to the response categories for the MGC items. Results: The fit statistics indicated that the 10 items belong together and can be summated for a total score. There was an overall absence of item order distortion between response categories. The Rasch model's expected category response values were compatible with item weights previously assigned. Conclusions: Our analysis suggests that: (1) the score can be summated to estimate an overall disease severity score; (2) the response options of the 10 items are not significantly distorted; and (3) the assigned weights of the response options are appropriate. Muscle Nerve 45: 820–825, 2012     

An IRT validation of the Affective Self Rating Scale

Background: The Affective Self Rating Scale (AS-18) is intended for the use in bipolar outpatients. It includes subscales for the rating of depressive and manic-type symptoms. It has previously been validated using methods from Classical Test Theory. Aims: The aim of this study was to evaluate the psychometric properties of the AS-18 when used at an outpatient clinic for patients with bipolar disorder at routine visits, and to analyze the potential for improvement of the scale. Methods: 231 patients with mainly bipolar I disorder doing ratings on routine visits at an affective disorder outpatient clinic were included. Ratings were analyzed using the Mokken non-parametric and the Rasch parametric model statistical methods. Results: In the Mokken analysis, both subscales of the AS-18 showed a strong ability to rank respondents according to their total score and all items contributed adequately to the measurement. In the Rasch model, there were no indications of disturbing influence from secondary dimensions in the subscales. The depression subscale had the capacity reliably to separate the sample in at most three levels and the mania subscale in two levels. The limited capacity to separate respondents can mainly be explained by a lack of items reflecting lower levels of depressive and manic symptoms. Conclusions: AS-18 has good basic psychometric properties for use of rating of symptoms in bipolar I patients at routine visits, but there is also room for improvement. Item Response Theory (IRT) methods are suitable tools for evaluation and construction of rating scales.     

Activity rating scales in adult muscle disease: What do they actually measure?

Introduction: There is a diversity of rating scales that assess function in muscle disease. Definitive knowledge of the content covered by these scales would help in making choices. Methods: We searched for activity rating scales used for muscle disease and assessed their content by linking scale items to the International Classification of Functioning, Disability and Health (ICF) and the muscle regions they cover. Results: Of the 119 scales found, 19 muscle disease–specific and 9 generic scales were prioritized for analysis. These 28 scales contained 457 items, from which 1145 concepts were identified and 160 were unique. Of these concepts, 97.8% could be linked to the ICF, most to the activities and participation domain (68.7%), followed by environmental factors (22.5%) and body functions (6.6%). Global muscle function was assessed most frequently, followed by lower and upper extremity function. Conclusions: This content comparison should allow for a better‐informed choice of activity rating scales for muscle disease. Muscle Nerve 50 : 14–23, 2014     

Attention Deficit Hyperactivity Disorder: A Rasch Analysis of the SWAN Rating Scale

The prevalence of attention-deficit/hyperactivity disorder (ADHD) has been estimated at 3–7% in the population. Children with this disorder are often characterized by symptoms of inattention and/or impulsivity and hyperactivity, which can significantly impact on many aspects of their behaviour and performance. This study investigated the characteristics of the SWAN Rating Scale and its discrimination of ADHD subtypes. This instrument was developed by Swanson and his colleagues and measures attentiveness and hyperactivity on a continuum, from attention problems to positive attention skills, using a seven-point scale of behaviour: “far below average” to “far above average”. The Australian Twin Attention-Deficit/Hyperactivity Disorder Study consists of questionnaire data collected from families in 1990/2007. The Rasch model was used to measure the characteristics of items from the SWAN Rating Scale; how well these items discriminated between those with and without ADHD. The prevalence of each subtype was found to be 5.3% for inattentive ADHD, 4.3% for hyperactive ADHD and 4.6% for combined ADHD. A total of 14.2% of the cohort appeared to have ADHD. While the inattentive items appeared to be consistent with each other in their measurement behaviour and response patterns, the hyperactive items were less consistent. Further, the combined subtype appeared to be an entirely different type, with unique features unlike the other two subtypes. Further work is needed to distinguish the diagnostic features of each subtype of ADHD.     

Rasch analysis of the Korean version of the VF‐14: calibration and unidimensionality

The VF‐14 is intended to measure visual functioning based on sets of hierarchically structured items. Rasch item response theory (IRT) provides a methodology to examine the hierarchical structure and unidimensionality of a survey instrument. The purpose of this study is to examine the scale properties (hierarchical structure and unidimensionality) of the VF‐14 using Rasch item response theory methodology in cataract patients (N = 80). The results of the study showed that the Korean version VF‐14 is unidimensional with a hierarchy from easy to hard items, as well as being valid and reliable in the terms of traditional psychometric analysis. The Korean version of the VF‐14 had some significant item spacings (three strata), and the rating‐scale categories were worded well. Therefore, the Korean version of the VF‐14 measures are appropriate for use in and further study of the visual limitations associated with cataract inpatients. Copyright © 2001 Whurr Publishers Ltd.     

Refining and validating the Social Interaction Anxiety Scale and the Social Phobia Scale

Background: The Social Interaction Anxiety Scale and Social Phobia Scale 6 are companion measures for assessing symptoms of social anxiety and social phobia. The scales have good reliability and validity across several samples, 3 , 6 however, exploratory and confirmatory factor analyses have yielded solutions comprising substantially different item content and factor structures. These discrepancies are likely the result of analyzing items from each scale separately or simultaneously. The current investigation sets out to assess items from those scales, both simultaneously and separately, using exploratory and confirmatory factor analyses in an effort to resolve the factor structure. Methods: Participants consisted of a clinical sample (n5353; 54% women) and an undergraduate sample (n5317; 75% women) who completed the Social Interaction Anxiety Scale and Social Phobia Scale, along with additional fear‐related measures to assess convergent and discriminant validity. Results: A three‐factor solution with a reduced set of items was found to be most stable, irrespective of whether the items from each scale are assessed together or separately. Items from the Social Interaction Anxiety Scale represented one factor, whereas items from the Social Phobia Scale represented two other factors. Conclusion: Initial support for scale and factor validity, along with implications and recommendations for future research, is provided. Depression and Anxiety, 2009. © 2009 Wiley‐Liss, Inc.     

Should the SCOPA‐COG be modified?: a Rasch analysis perspective

Background and purpose: The SCales for Outcomes in PArkinson’s disease‐Cognition (SCOPA‐COG) is a specific measure of cognitive function for Parkinson’s disease (PD) patients. Previous studies, under the frame of the classic test theory, indicate satisfactory psychometric properties. The Rasch model, an item response theory approach, provides new information about the scale, as well as results in a linear scale. This study aims at analysing the SCOPA‐COG according to the Rasch model and, on the basis of results, suggesting modification to the SCOPA‐COG. Method: Fit to the Rasch model was analysed using a sample of 384 PD patients. Results: A good fit was obtained after rescoring for disordered thresholds. The person separation index, a reliability measure, was 0.83. Differential item functioning was observed by age for three items and by gender for one item. Conclusions: The SCOPA‐COG is a unidimensional measure of global cognitive function in PD patients, with good scale targeting and no empirical evidence for use of the subscale scores. Its adequate reliability and internal construct validity were supported. The SCOPA‐COG, with the proposed scoring scheme, generates true linear interval scores.     

Comparison of functional activity performance in normally developing children and children with cerebral palsy

OBJECTIVE: To compare the pattern of self-care performance in normal children and children with cerebral palsy (CP). METHOD: 142 normal children and 33 children with CP were evaluated by 22 items from the self-care scale of the PEDI functional test. Rasch methodology transformed scores into interval measures of difficulty from 0 to 100 (logit). Spearman rank correlation coefficient compared the order of logits in the two groups. RESULTS: Eleven items showed significant differences in the logit values received. Out of these, 7 items showed relative difficulty values greater in the group of children with CP and 4 items showed relative difficulty values greater among normal children. A significant correlation was observed in the order of the 22 items displayed in the two interval scales. CONCLUSION: The results suggest that the development of self-care functional activities may be influenced by the presence of CP. These results may support assessment and intervention strategies for children with neuromotor disorders.     

Construction and validation of the chronic acquired polyneuropathy patient‐reported index (CAP‐PRI): A disease‐specific,health‐related quality‐of‐life instrument

Introduction: Generic health‐related quality‐of‐life (HRQOL) patient‐reported outcome measures have been used in patients with chronic immune‐mediated polyneuropathies. We have created a disease‐specific HRQOL instrument. Methods: The chronic acquired polyneuropathy patient‐reported index (CAP‐PRI) was developed and validated in multiple steps. Items were initially generated through patient and specialist input. The performance of the preliminary 20 items was analyzed via a prospective, 5‐center study involving chronic immune‐mediated polyneuropathy patients. Results: Data analysis suggested modification to a 15‐item scale with 3 response categories rather than 5. The final CAP‐PRI was validated in another prospective, 5‐center study. The CAP‐PRI appeared to be a unidimensional outcome measure that fit the Rasch model in our multicenter cohort. It correlated appropriately with outcome measures commonly used in this patient population. Conclusions: The CAP‐PRI is a simple disease‐specific HRQOL measure that appears to be useful for clinical care and possibly also for clinical trials. Muscle Nerve 54 : 9–17, 2016