Interpreting results from Rasch analysis 2. Advanced model applications and the data-model fit assessment

Psychometric evaluation of the wijma delivery expectancy/experience questionnaire for pregnant women in Taiwan

BACKGROUND

Fear of childbirth profoundly affects women's ability to cope during pregnancy and influences birth outcomes. In Taiwan, there's a lack of validated tools for assessing childbirth fear.

OBJECTIVE

To evaluate the psychometric properties of the Taiwanese version of the Wijma delivery expectancy/experience questionnaire version A (WDEQ-A) in pregnant women.

METHODS

We conducted a cross-sectional study with pregnant women, using the WDEQ-A and a Visual Analogue Scale to assess childbirth fear. We employed the Edinburgh Postpartum Depression Scale, State-Trait Anxiety Inventory, and Mindful Attention Awareness Scale to measure depressive symptoms, anxiety, and mindfulness, respectively. We evaluated internal consistency reliability and construct validity using exploratory factor analysis and Rasch analysis. Pearson correlations measured the association between childbirth fear and psychological variables. The receiver operating characteristic curve was used for determining the sensitivity and specificity indices.

RESULTS

The WDEQ-A demonstrated excellent reliability (Cronbach's alpha = 0.93) and revealed a three-factor structure, including fear, lack of positive anticipation, and isolation. Rasch analysis supported the dimensionality of each of the three revised factors. The total score significantly correlated with depression (r = 0.56), anxiety (r = 0.19), and mindfulness (r = -0.40) (Ps < 0.001). The optimal cut-off for WDEQ-A was established at 55.5, with a sensitivity of 83 % and a specificity of 65 %.

CONCLUSIONS

The Taiwanese version of WDEQ-A proved to be a reliable and valid tool for assessing fear of childbirth in pregnant women, demonstrating excellent psychometric properties.

IMPLICATION FOR PRACTICE

These findings can assist midwives in Taiwan in identifying and promptly providing effective strategies for women experiencing a high fear of childbirth.

The psychometric properties of the Pearlin Mastery Scale in persons living with restless legs syndrome

INTRODUCTION

Restless Legs Syndrome (RLS) is a neurological disorder characterized by an urge to move arms and legs, commonly combined with distress, pain and motor restlessness. It can cause fragmented sleep, daytime symptoms, and decreased quality of life. Pharmacological treatment can suppress symptoms, but not cure. When challenged with illness, people may turn to their inner psychological resources such as self-esteem and mastery. The Pearlin Mastery scale was developed to study stress and coping, is commonly used in people with chronic illnesses, however, not yet validated in people with RLS.

AIM

The aim was to test reliability and construct validity of the Pearlin Mastery Scale in persons with RLS.

METHODS

A cross-sectional postal survey including the Pearlin Mastery Scale, Restless Legs Syndrome-6 Scale, Pittsburgh Sleep Quality Index and Patient Health Questionnaire was sent out to members (n = 1500) of the national RLS association and 788 (52.5%) agreed to participate. Data were analyzed using classical test theory, Confirmatory factor analysis and Rasch measurement theory analysis. Hypothesis testing for construct validity was done by bivariate correlation analyses.

RESULTS

Most respondents were women (65%), retired (71%) and had a mean-age of 70.8 years (SD 11.4). The 7-item version of the Pearlin Mastery Scale showed poor fit to the one factor model. After omitting the two positively worded items (i.e., item 4 and 6), the 5-item version was found to be unidimensional, with satisfactory internal consistency. However, all items showed considerable ceiling effects. No measurement variance was seen regarding age-groups or sex. Higher level of mastery was moderately correlated with less depressive symptoms but only weakly correlated to RLS-related sleep problems.

CONCLUSION

The 5-item version of the Pearlin Mastery Scale is suggested to be used in persons with RLS due to its acceptable psychometric properties. The instrument could be applied as an outcome measure for behavioral change interventions aiming to support mastery in RLS.

Generalisability of the Barthel Index and the Functional Independence Measure: robustness of disability measures to Differential Item Functioning

PURPOSE

Differential Item Functioning (DIF), an item malfunctioning, causes Differential Test Functioning (DTF), thus biasing questionnaire measures. The current study evaluates the relationship between DIF and DTF for the Barthel Index and the Functional Independence Measure, likely the most used disability measures. The aim is to understand under which conditions DIF can be ignored as its DTF is negligible.

METHODS

A simulation study was run. Disability measures were obtained for the Barthel Index and FIM motor domain using Rasch analysis with previously published item calibrations. Several DIF scenarios have been assessed. DTF was tolerable if ≤0.50 logits.

RESULTS

Simulations showed that the larger the DIF, the larger the DTF and that, keeping the overall DIF constant, the total number of items with DIF does not affect DTF. DIF of the items with the lowest or highest calibrations is the most dangerous. The DIF of central items should be so massive to matter in DTF terms that it is unlikely to happen in practice. The FIM robustness to DIF is better than that of the Barthel Index.

CONCLUSIONS

The FIM and the Barthel Index show remarkable robustness to DIF. Thanks to this feature, sample invariant, generalisable disability measures are available.

A Multi-Sample Comparison and Rasch Analysis of the Evaluation of Children's Listening and Processing Skills Questionnaire

OBJECTIVES

Assessing listening difficulties and associated complaints can be challenging. Often, measures of peripheral auditory functions are within normal ranges, making clinicians feel unsure about proper management strategies. The range and nature of observed or experienced difficulties might be better captured using a qualitative measure. The Evaluation of Children's Listening and Processing Skills (ECLiPS) questionnaire was designed to broadly profile the auditory and cognitive problems often present in children with listening difficulties. This 38-item questionnaire was initially standardized in British children aged 6 to 11 years, was subsequently modified for use with North-American children, and was recently translated into Flemish-Dutch. This study aimed to compare typical scores of the Flemish version with the UK and US versions, and to evaluate and compare its psychometric quality based on Rasch analysis.

DESIGN

We selected 112 Flemish children aged 6 to 11 years with verified normal hearing and typical development, and asked two caregivers of every child to fill out the ECLiPS. Data from two comparator samples were analyzed, including responses for 71 North-American children and 650 British children. Typical values for ECLiPS factors and aggregates were determined as a function of age and gender, and meaningful differences across samples were analyzed. Rasch analyses were performed to evaluate whether ECLiPS response categories work as intended, and whether item scores fit a linear equal interval measurement scale that works the same way for everyone. Item and person metrics were derived, including separation and reliability indices. We investigated whether items function similarly across linguistically and culturally different samples.

RESULTS

ECLiPS scores were relatively invariant to age. Girls obtained higher scores compared with boys, mainly for items related to memory and attention, and pragmatic and social skills. Across ECLiPS versions, the most pronounced differences were found for items probing social skills. With respect to its psychometric quality, ECLiPS response categories work as intended, and ECLiPS items were found to fit the Rasch measurement scale. Cultural differences in responses were noted for some items, belonging to different factors. Item separation and reliability indices generally pointed toward sufficient variation in item difficulty. In general, person separation (and reliability) metrics, quantifying the instrument's ability to distinguish between poor and strong performers (in a reproducible manner), were low. This is expected from samples of typically developing children with homogeneous and high levels of listening ability.

CONCLUSIONS

Across the languages assessed here, the ECLiPS caregiver questionnaire was verified to be a psychometrically valid qualitative measure to assess listening and processing skills, which can be used to support the assessment and management of elementary school children referred with LiD.

In Myotonic Dystrophy Type 1 Head Repositioning Errors Suggest Impaired Cervical Proprioception

Background: Myotonic dystrophy type 1 (DM1) is a rare multisystemic genetic disorder with motor hallmarks of myotonia, muscle weakness and wasting. DM1 patients have an increased risk of falling of multifactorial origin, and proprioceptive and vestibular deficits can contribute to this risk. Abnormalities of muscle spindles in DM1 have been known for years. This observational cross-sectional study was based on the hypothesis of impaired cervical proprioception caused by alterations in the neck spindles. Methods: Head position sense was measured in 16 DM1 patients and 16 age- and gender-matched controls. A head-to-target repositioning test was requested from blindfolded participants. Their head was passively rotated approximately 30° leftward or rightward and flexed or extended approximately 25°. Participants had to replicate the imposed positions. An optoelectronic system was adopted to measure the angular differences between the reproduced and the imposed positions (joint position error, JPE, °) concerning the intended (sagittal, horizontal) and unintended (including the frontal) planar projections. In DM1 patients, JPEs were correlated with clinical and balance measures. Static balance in DM1 patients was assessed through dynamic posturography. Results: The accuracy and precision of head repositioning in the intended sagittal and horizontal error components did not differ between DM1 and controls. On the contrary, DM1 patients showed unintended side-bending to the left and the right: the mean [95%CI] of frontal JPE was -1.29° [-1.99°, -0.60°] for left rotation and 0.98° [0.28°, 1.67°] for right rotation. The frontal JPE of controls did not differ significantly from 0° (left rotation: 0.17° [-0.53°, 0.87°]; right rotation: -0.22° [-0.91°, 0.48°]). Frontal JPE differed between left and right rotation trials (p < 0.001) only in DM1 patients. No correlation was found between JPEs and measures from dynamic posturography and clinical scales. Conclusions: Lateral head bending associated with head rotation may reflect a latent impairment of neck proprioception in DM1 patients.

Physical and Rehabilitation Medicine: say relational or functional, not holistic

Modern medicine tends to privilege disciplines promising "objective" laws governing body parts (from molecules to organs). Studies on a person's illness and disability are (apparently) confined to "subjectivity." The Specialty of Physical and Rehabilitation Medicine is often regarded as a humanitarian approach, belonging at best to the family of "soft," "qualitative," or "quasi-experimental" sciences. This specialty often claims specificity by labelling itself as "functional" and "holistic." However, it is shown here that the former term is acceptable, yet redundant, and the second misleading. When human behaviors and perceptions are at stake, "function" indicates a person's relationship with the outer world (already tackled by the definitional term "physical" from the Greek "physis"). The word "holistic" emphasizes mind-body unity and person-environment interdependence but, in current usage, overshadows the complementary need for an analytic, experimental approach to any function. Medicine aims at fighting disease and disability in single persons. This endeavor requires knowing body parts and mechanisms and understanding how interventions on "parts" affect the "whole." This understanding rests on the experimental method. For instance, returning to a given societal role (participation) may require restoration of walking (activity), which may require reinforcement of weakened muscular groups (impairment). Working only on holistic bio-psycho-social "wholes" may miss the therapeutic mission of medicine.

Interpreting results from Rasch analysis 1. The "most likely" measures coming from the model

Purpose: The present article summarises the characteristics of Rasch's theory, providing an original metrological model for persons' measurements. Properties describing the person "as a whole" are key outcome variables in Medicine. This is particularly true in Physical and Rehabilitation Medicine, targeting the person's interaction with the outer world. Such variables include independence, pain, fatigue, balance, and the like. These variables can only be observed through behaviours of various complexity, deemed representative of a given "latent" person's property. So how to infer its "quantity"? Usually, behaviours (items) are scored ordinally, and their "raw" scores are summed across item lists (questionnaires). The limits and flaws of scores (i.e., multidimensionality, non-linearity) are well known, yet they still dominate the measurement in Medicine.Conclusions: Through Rasch's theory and statistical analysis, scores are transformed and tested for their capacity to respect fundamental measurement axioms. Rasch analysis returns the linear measure of the person's property ("ability") and the item's calibrations ("difficulty"), concealed by the raw scores. The difference between a person's ability and item difficulty determines the probability that a "pass" response is observed. The discrepancy between observed scores and the ideal measures (i.e., the residual) invites diagnostic reasoning. In a companion article, advanced applications of Rasch modelling are illustrated. Implications for rehabilitationQuestionnaires' ordinal scores are poor approximations of measures. The Rasch analysis turns questionnaires' scores into interval measures, provided that its assumptions are respected.Thanks to the Rasch analysis, accurate measures of independence, pain, fatigue, cognitive capacities and other whole person's variables of paramount importance in rehabilitation are available.The current work is addressed to rehabilitation professionals looking for an introduction to interpreting published results based on Rasch analysis.The first of a series of two, the present article illustrates the most common graphic and numeric outputs found in published papers presenting the Rasch analysis of questionnaires.

Fatigue may improve equally after balance and endurance training in multiple sclerosis: a randomised, crossover clinical trial

Introduction

Fatigue and poor balance are frequent and severe problems in multiple sclerosis (MS) that may interact. Endurance training is known to be effective on fatigue. This study aims to test if balance training is more effective against MS fatigue.

Methods

A randomised crossover trial was run, recruiting 31 MS people (21 women; median age: 46 years, range: 30-64; median EDSS: 4, range: 2.5-5). Participants received balance and endurance training alternately (15 one-to-one sessions, 5 days/week) and were assessed before (T0), after (T1), and 30 days after treatment ended (T2). The Modified Fatigue Impact Scale (MFIS) with scores linearised through Rasch analysis was the primary outcome (the lower the measure, the better the condition, i.e., the lower the fatigue symptoms). The Equiscale balance scale and posturography (EquiTest) were used to assess balance. Linear mixed-effects models with ANOVA were used for significance testing.

Results

Thirteen participants had no carryover effect and were included in the primary analysis. Fatigue significantly changed across the three time points (F2,58 = 16.0; p < 0.001), but no difference across treatments was found. Altogether, both treatments significantly improved the MFIS measure at T1 (95%CI: -1.24 logits; mean: -1.67 to -0.81 logits) and T2 (95%CI: -1.04; mean: -1.49 to -0.60) compared to T0 (95%CI: -0.51; mean: -0.95 to -0.08; p ≤ 0.001). Equiscale and posturography highlighted balance improvement after balance training but not after endurance training.

Conclusion

Balance and endurance training could similarly reduce fatigue in MS patients in the short term. However, only balance training also improved balance in MS.

Pay attention: you can fall! The Mini-BESTest scale and the turning duration of the TUG test provide valid balance measures in neurological patients: a prospective study with falls as the balance criterion

Background

Balance, i.e., the ability not to fall, is often poor in neurological patients and this impairment increases their risk of falling. The Mini-Balance Evaluation System Test (Mini-BESTest), a rating scale, the Timed Up and Go (TUG) test, and gait measures are commonly used to quantify balance. This study assesses the criterion validity of these measures as balance measures.

Methods

The probability of being a faller within nine months was used as the balance criterion. The Mini-BESTest, TUG (instrumented with inertial sensors), and walking test were administered before and after inpatient rehabilitation. Multiple and LASSO logistic regressions were used for the analysis. The diagnostic accuracy of the model was assessed with the area under the curve (AUC) of the receiver operating characteristic curve. Mobility measure validity was compared with the Akaike Information Criterion (AIC).

Results

Two hundred and fourteen neurological patients (stroke, peripheral neuropathy, or parkinsonism) were recruited. In total, 82 patients fell at least once in the nine-month follow-up. The Mini-BESTest (AUC = 0.69; 95%CI: 0.62-0.76), the duration of the TUG turning phase (AUC = 0.69; 0.62-0.76), and other TUG measures were significant faller predictors in regression models. However, only the turning duration (AIC = 274.0) and Mini-BESTest (AIC = 276.1) substantially improved the prediction of a baseline model, which only included fall risk factors from the medical history (AIC = 281.7). The LASSO procedure selected gender, disease chronicity, urinary incontinence, the Mini-BESTest, and turning duration as optimal faller predictors.

Conclusion

The TUG turning duration and the Mini-BESTest predict the chance of being a faller. Their criterion validity as balance measures in neurological patients is substantial.

Sensory Modulation Abilities in Healthy Preterm-Born Children: An Observational Study Using the Sensory Processing and Self-Regulation Checklist (SPSRC)

This study aimed to investigate prematurity as a risk factor for sensory processing disorders, using the Italian Version of Sensory Processing and Self-Regulation Checklist (SPSRC-IT), based on a sample of healthy Italian children born preterm in comparison with a sample of typical full-term children. Two groups of caregivers of Italian healthy preschooler children were recruited. The first group comprised 37 caregivers of full-term children (FT), while the second group consisted of 37 caregivers of preterm children (PT) (gestational age < 37 weeks). Significant differences between the groups in several subsections and factors of the SPSRC-IT were found, specifically in the Physiological Conditions section, in the Gustatory and Olfactory Sense section, in the Vestibular Sense section, and in the Proprioceptive Sense section, with lower scores in the PT group. Moreover, children born at a lower gestational age or with lower weights had a higher risk of dysfunctions in processing gustatory and olfactory, vestibular, and proprioceptive stimuli. In conclusion, the SPSRC-IT suggested a potential link between prematurity and challenges in the development of sensory processing and self-regulation skills, especially in children with a very low birth weight and very low gestational age.

Reversed Mirror Therapy (REMIT) after Stroke-A Proof-of-Concept Study

In mirror training (MIT), stroke patients strive to move their hands while looking at the reflected image of the unaffected one. The recruitment of the mirror neurons and visual-proprioceptive conflict are expected to facilitate the paretic voluntary movement. Here, a reversed MIT (REMIT) is presented, which requires moving hands while looking at the reflected image of the paretic one, giving the illusion of being unable to move the unimpaired hand. This study compares MIT and REMIT on post-stroke upper-limb recovery to gain clues on the mechanism of action of mirror therapies. Eight chronic stroke patients underwent two weeks of MIT and REMIT (five sessions each) in a crossover design. Upper-limb Fugl-Meyer, Box and Block and handgrip strength tests were administered at baseline and treatments end. The strength of the mirror illusion was evaluated after each session. MIT induced a larger illusory effect. The Fugl-Meyer score improved to the same extent after both treatments. No changes occurred in the Box and Block and the handgrip tests. REMIT and MIT were equally effective on upper-limb dexterity, challenging the exclusive role of mirror neurons. Contrasting learned nonuse through an intersensory conflict might provide the rationale for both forms of mirror-based rehabilitation after stroke.

Differential Item Functioning of the Mini-BESTest Balance Measure: A Rasch Analysis Study

The Mini-Balance Evaluation Systems Test (Mini-BESTest), a 14-item scale, has high content validity for balance assessment. This study further examines the construct validity of the Mini-BESTest with an emphasis on its measurement invariance. The Mini-BESTest was administered to 292 neurological patients in two sessions (before and after rehabilitation) and evaluated with the Rasch analysis (Many-Facet Rating Scale Model: persons, items, sessions). Categories' order and fit to the model were assessed. Next, maps, dimensionality, and differential item functioning (DIF) were examined for construct validity evaluation. DIF was inspected for several clinically important variables, including session, diagnosis, and assistive devices. Mini-BESTest items had ordered categories and fitted the Rasch model. The item map did not flag severe construct underrepresentation. The dimensionality analysis showed that another variable extraneous to balance affected the score of a few items. However, this multidimensionality had only a modest impact on measures. Session did not cause DIF. DIF for assistive devices affected six items and caused a severe measurement artefact. The measurement artefact caused by DIF for diagnosis was negligible. The Mini-BESTest returns interval measures with robust construct validity and measurement invariance. However, caution should be used when comparing Mini-BESTest measures obtained with and without assistive devices.