key: cord-0819751-jl894pzy authors: Deng, W.; Carpentier, S.; Blackwood, J.; Van de Winckel, A. title: Rasch Validation of the Warwick-Edinburgh Mental Well-being Scale (WEMWBS) in Community-Dwelling Adults and in Adults with Stroke in the US date: 2022-04-19 journal: nan DOI: 10.1101/2022.04.18.22274001 sha: 80df1315df3bd0d7b56eb103be9377f2249cd11b doc_id: 819751 cord_uid: jl894pzy Background: With the recent ongoing global COVID-19 pandemic and political divide in the United States (US), there is an urgent need to address the soaring mental well-being problems and to promote positive well-being. The Warwick-Edinburgh Mental Well-being Scale (WEMWBS) measures the positive aspects of mental health. Previous studies confirmed its construct validity, reliability, and unidimensionality with confirmatory factor analysis. Four studies have performed a Rasch analysis on the WEMWBS, but none of them tested adults in the US. The goals of our study are to use Rasch analysis to validate the WEMWBS in the general US population and in adults with stroke. Methods: We recruited community-dwelling adults and adults with chronic stroke with upper limb hemiplegia or hemiparesis. We used the Rasch Unidimensional Measurement Model (RUMM) 2030 software to evaluate item and person fit, targeting, person separation reliability (PSR), and differential item functioning (DIF) for sample sizes of at least 200 persons in each subgroup. Results: After deleting two items, the WEMBS analyzed in our 553 community-dwelling adults (average age 51.22 +/- 17.18 years; 358 women) showed an excellent PSR=0.91 as well as person and item fit, but the items are too easy for this population (person mean location=2.17 +/- 2.00). There was no DIF for sex, mental health, or practicing breathing exercises. In the 37 adults with chronic stroke (average age 58 +/- 13; 11 women) the WEMWBS had a good item and person fit, and PSR=0.92, but the items were too easy for this group as well (person mean location=3.13 +/- 2.00). Conclusions: The WEMWBS had good item and person fit but the targeting is off when used in community-dwelling adults and adults with stroke in the US. Adding more difficult items might improve the targeting and capture a broader range of positive mental wellbeing in both populations. Our pilot data in adults with stroke needs to be confirmed in a larger sample size. Background 54 In recent years, the global COVID-19 pandemic has resulted in overworked health care workers, 55 and many adults facing serious health problems, death of loved ones, and fear of losing their 56 job. [1] Coupled with a rise in violence caused by a political divide, the United States (US) has 57 seen a 10% increase in prevalence of adults with serious psychological distress in 2020 58 compared to 2018. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 19, 2022. ; https://doi.org/10.1101/2022.04.18.22274001 doi: medRxiv preprint Aside from mental health problems in the general adult population, adults who experience a 98 stroke are particularly vulnerable to depression, with approximately 30% of stroke survivors 99 experiencing post-stroke depression at any given time. [13] Post-stroke depression is related to 100 poor rehabilitation outcomes. In contrast, an increase in positive emotions over a 3-month period 101 post-stroke is associated with an increased likelihood of functional recovery, which may lead to 102 improved quality of life. [14, 15] Therefore, it is important to measure positive mental well-being 103 in people with stroke and to assess whether WEMWBS would be a good measure for this 104 population, but this has not yet been investigated. 105 106 Therefore, our first aim of this study is to assess the structural validity of the WEMWBS with 107 Rasch in community-dwelling adults in the US. We will compare our findings with prior Rasch 108 results in other countries. Our second aim is to perform a pilot Rasch validation on the 109 WEMWBS in adults with chronic stroke. For this cross-sectional study, we recruited participants at the Minnesota State Fair and Highland 115 Fest and through volunteer sampling using research fliers and study postings on relevant 116 websites. We also emailed the flier to volunteers who expressed interest in research from the 117 Brain Body Mind Lab at the University of Minnesota. Recruitment occurred from September 27, 118 2017 till August 12, 2020. For both community-dwelling adults and adults with stroke, we 119 included adults between 18-99 years of age, English speaking, and able to consent. Additionally, 120 adults with stroke were included if they had an ischemic or hemorrhagic stroke and were 121 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The Warwick questionnaire covers positive aspects of mental health. All 14 items have a scoring 137 range from "0-None of the time" to "4-All of the time". A higher score on each item indicates a 138 more positive attitude towards life. We collected demographic information, and whether 139 participants currently practiced mindfulness, breathing exercises, or body awareness exercises 140 (e.g., Yoga, Qigong, Pilates). We inquired whether they had current pain conditions or current 141 mental health conditions. 142 143 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Component Analysis of Residuals (PCAR), which refers to the extent to which covariance in the 162 residuals is random and not explained underlying constructs than the one that is being 163 measured. [10, 22] In that case, the expected eigenvalue is less than 2, and the percent variance 164 explained by the first component is less than 10%. If those criteria are not met, then dependent t-165 tests between the 2 subsets of items with positive and negative loadings on the first residual 166 component are performed. We would confirm unidimensionality if less than 5% of these tests are 167 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. We recruited 553 community-dwelling adults and 37 adults with stroke. The characteristics of 175 the demographic and clinical information of all participants is presented in Table 1 . 176 177 The iteration analysis displays the step-by-step approach taken for the Rasch analysis 179 (Additional file 1). The main results are described below. 180 For our first analysis in community-dwelling Americans, none of the 14 items displayed 181 disordered thresholds. Two items were misfitting: item 1 "I have been feeling optimistic about 182 the future" and item 5 "I have had energy to spare." After deleting items 1 and 5, all items fit the 183 model and only 2.71% of persons were misfitting. The hierarchy of the item difficulty is 184 presented in Figure 1 , with the easiest items starting at the top and the hardest items at the 185 bottom. The item logit location and fit statistics are presented in Table 2 . There was no floor or 186 ceiling effect, but the person mean location ± standard deviation was 2.17±2.00 logits, meaning 187 that the items were too easy for this population (Figure 2) too easy (person mean location 1.88±1.71). We therefore do not recommend using the 7-items 214 scale for clinical use. We recommend that the targeting first be solved before it can be used in 215 the clinic or for research and therefore, we do not provide a revised scoring sheet or score-to-216 measure table for the 12-item revised scale. 217 218 Our pilot Rasch Measurement Theory analysis in adults with chronic stroke (n=37) revealed that 219 item 7 had reversed thresholds. After rescoring item 7 to scoring categories [00123], all items fit 220 (Table 3) and there were no more reversed thresholds (Figure 3) . Only 2.70% of the people were 221 misfitting. There was no floor or ceiling effect, and the WEMWBS had an excellent PSR of 0.92. 222 However, the person mean location was 3.13±2.00 (Figure 4) . Even though these results need to 223 be validated in a larger study, they seem to indicate that those items are also too easy for adults 224 with chronic stroke and very few participants choose the lowest category of "None of the time" 225 or "Rarely" (Table 4) (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 19, 2022. The WEMWBS demonstrated good item fit and person fit in a US population of community-297 dwelling adults and adults with stroke. However, the items are too easy, which is a consistent 298 finding across the majority of WEMWBS Rasch studies performed in different countries. Thus, 299 including more difficult items in a next iteration of the scale could help solve the targeting. 300 Finally, further studies with larger sample sizes should be performed in adults with stroke, and 301 perhaps include adults with acute and subacute stroke, to validate our preliminary findings. 302 303 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. There is no financial competing interest in this study. We appreciate all the participants who have participated in the study, and the research volunteers 347 who have helped with data collection. Our profound gratitude goes to Marc Noël for the critical 348 review of the manuscript. 349 350 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 19, 2022. ; https://doi.org/10.1101/2022.04.18.22274001 doi: medRxiv preprint Positive psychology in a pandemic: buffering, 362 bolstering, and building mental health Psychological Distress and 364 Loneliness Reported by US Adults On happiness and human potentials: a review of research on 366 hedonic and eudaimonic well-being Construction and Validation of a Measure of Integrative 368 Well-Being in Seven Languages: The Pemberton Happiness Index.(PHI); Copyright: 369 Creative Commons License The Warwick-Edinburgh Mental Well-being Scale (WEMWBS): 372 development and UK validation Internal 374 construct validity of the Warwick-Edinburgh Mental Well-being Scale (WEMWBS): a 375 Rasch analysis using data from the Scottish Health Education Population Survey Mental Well-Being: A Validation of a Rasch-Derived Version of the Warwick-Edinburgh 379 Mental Well-Being Scale Edinburgh_Mental_Wellbeing_Scale_Rasch_Analysis_in_the_Indonesian_Context/links/60 383 cbf3afa6fdcc01d47dd288/Internal-Validation-of-the-Warwick Wellbeing-Scale-Rasch-Analysis-in-the-Indonesian-Context.pdf Mental Well-being Scale (WEMWBS) in the UK veterinary profession The Rasch measurement model in rheumatology: what is 389 it and why use it? When should it be applied, and what should one look for in a Rasch 390 paper? Reporting Guideline for RULER: Rasch Reporting Guideline for Rehabilitation 393 Research -Explanation & Elaboration manuscript Rasch Reporting Guideline for Rehabilitation Research (RULER): The RULER Statement Part I: frequency of depression after stroke: an updated 399 systematic review and meta-analysis of observational studies Change in positive emotion and 401 recovery of functional status following stroke Associations 403 between positive emotion and recovery of functional status following stroke Mini-mental state": a practical method for 406 grading the cognitive state of patients for the clinician Aphasia Rapid Test: an NIHSS-like aphasia test Comprehensive assessment of gesture production: a new 411 test of upper limb apraxia (TULIA) A new brief instrument for assessing decisional capacity for clinical 414 research Psychometric evaluation of an interview-administered 416 version of the WHOQOL-BREF questionnaire for use in a cross-sectional study of a rural 417 district in Bangladesh: an application of Rasch analysis Reliability and separation of measures Individual-patient monitoring in clinical practice: are 425 available health status surveys adequate? Development and validation of the activity significance personal evaluation 428 (ASPEn) scale Foundations of behavioral research Cross cultural 432 evaluation of the Warwick-Edinburgh Mental Well-being Scale (WEMWBS) --a mixed 433 methods study Well-being Scale (WEMWBS) in Northern Ireland