Limits...
Rasch Analysis of the Adult Strabismus Quality of Life Questionnaire (AS-20) among Chinese Adult Patients with Strabismus.

Wang Z, Zhou J, Luo X, Xu Y, She X, Chen L, Yin H, Wang X - PLoS ONE (2015)

Bottom Line: The overall AS-20 did not demonstrate unidimensional; however, it was achieved separately in the two Rasch-revised subscales: the psychosocial subscale (11 items) and the function subscale (9 items).The features of good targeting, optimal item infit and outfit, and no notable local dependence were found for each of the subscales.The rating scale was appropriate for the psychosocial subscale but a reduction to four response categories was required for the function subscale.

View Article: PubMed Central - PubMed

Affiliation: School of Nursing, Third Military Medical University, Chongqing, China.

ABSTRACT

Background: The impact of strabismus on visual function, self-image, self-esteem, and social interactions decrease health-related quality of life (HRQoL).The purpose of this study was to evaluate and refine the adult strabismus quality of life questionnaire (AS-20) by using Rasch analysis among Chinese adult patients with strabismus.

Methods: We evaluated the fitness of the AS-20 with Rasch model in Chinese population by assessing unidimensionality, infit and outfit, person and item separation index and reliability, response ordering, targeting and differential item functioning (DIF).

Results: The overall AS-20 did not demonstrate unidimensional; however, it was achieved separately in the two Rasch-revised subscales: the psychosocial subscale (11 items) and the function subscale (9 items). The features of good targeting, optimal item infit and outfit, and no notable local dependence were found for each of the subscales. The rating scale was appropriate for the psychosocial subscale but a reduction to four response categories was required for the function subscale. No significant DIF were revealed for any demographic and clinical factors (e.g., age, gender, and strabismus types).

Conclusion: The AS-20 was demonstrated by Rasch analysis to be a rigorous instrument for measuring health-related quality of life in Chinese strabismus patents if some revisions were made regarding the subscale construct and response options.

No MeSH data available.


Related in: MedlinePlus

The category probability curves for (A) 5-response psychosocial subscale (representative item #1), (B) 5-response function subscale (representative item #11), (C) 4-response function subscale after the ‘rarely’ and ‘never’ response options were combined, (D) 4-response function subscale after the ‘rarely’ and ‘sometimes’ response options were combined.All curves at the extreme left represents ‘always’ and at the extreme right represents ‘never’. Response categories were properly oriented and distributed for all items in the psychosocial subscale (A). Using 5 response options for the function subscale, it was evident that the ‘rarely’ response was underutilized (B). When the ‘rarely’ option was combined with either the ‘never’ or the ‘sometimes’ option in the function subscale, response categories were properly oriented and distributed (C & D).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4636299&req=5

pone.0142188.g003: The category probability curves for (A) 5-response psychosocial subscale (representative item #1), (B) 5-response function subscale (representative item #11), (C) 4-response function subscale after the ‘rarely’ and ‘never’ response options were combined, (D) 4-response function subscale after the ‘rarely’ and ‘sometimes’ response options were combined.All curves at the extreme left represents ‘always’ and at the extreme right represents ‘never’. Response categories were properly oriented and distributed for all items in the psychosocial subscale (A). Using 5 response options for the function subscale, it was evident that the ‘rarely’ response was underutilized (B). When the ‘rarely’ option was combined with either the ‘never’ or the ‘sometimes’ option in the function subscale, response categories were properly oriented and distributed (C & D).

Mentions: The revised psychosocial subscale consisted of the item 1–10 and the item 17. The person separation index (PSI) was 2.70, with a reliability value of 0.88. The item separation index (ISI) was 7.16, with the reliability of 0.98 (Table 4). The difference between the person and item means was less than 1.0 logits (0.45 ± 1.27 logits for mean person vs. 0.00 ± 0.57 logits for mean item; Fig 2A), which suggested good targeting that the item difficulty matched well with the participants’ HRQoL. No disordered response categories were evident (Fig 3A), indicating each category properly represented stepwise increase in severity and the difference between categories on severity could be correctly endorsed by the respondents. Therefore the five-level Likert scaling was retained.


Rasch Analysis of the Adult Strabismus Quality of Life Questionnaire (AS-20) among Chinese Adult Patients with Strabismus.

Wang Z, Zhou J, Luo X, Xu Y, She X, Chen L, Yin H, Wang X - PLoS ONE (2015)

The category probability curves for (A) 5-response psychosocial subscale (representative item #1), (B) 5-response function subscale (representative item #11), (C) 4-response function subscale after the ‘rarely’ and ‘never’ response options were combined, (D) 4-response function subscale after the ‘rarely’ and ‘sometimes’ response options were combined.All curves at the extreme left represents ‘always’ and at the extreme right represents ‘never’. Response categories were properly oriented and distributed for all items in the psychosocial subscale (A). Using 5 response options for the function subscale, it was evident that the ‘rarely’ response was underutilized (B). When the ‘rarely’ option was combined with either the ‘never’ or the ‘sometimes’ option in the function subscale, response categories were properly oriented and distributed (C & D).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4636299&req=5

pone.0142188.g003: The category probability curves for (A) 5-response psychosocial subscale (representative item #1), (B) 5-response function subscale (representative item #11), (C) 4-response function subscale after the ‘rarely’ and ‘never’ response options were combined, (D) 4-response function subscale after the ‘rarely’ and ‘sometimes’ response options were combined.All curves at the extreme left represents ‘always’ and at the extreme right represents ‘never’. Response categories were properly oriented and distributed for all items in the psychosocial subscale (A). Using 5 response options for the function subscale, it was evident that the ‘rarely’ response was underutilized (B). When the ‘rarely’ option was combined with either the ‘never’ or the ‘sometimes’ option in the function subscale, response categories were properly oriented and distributed (C & D).
Mentions: The revised psychosocial subscale consisted of the item 1–10 and the item 17. The person separation index (PSI) was 2.70, with a reliability value of 0.88. The item separation index (ISI) was 7.16, with the reliability of 0.98 (Table 4). The difference between the person and item means was less than 1.0 logits (0.45 ± 1.27 logits for mean person vs. 0.00 ± 0.57 logits for mean item; Fig 2A), which suggested good targeting that the item difficulty matched well with the participants’ HRQoL. No disordered response categories were evident (Fig 3A), indicating each category properly represented stepwise increase in severity and the difference between categories on severity could be correctly endorsed by the respondents. Therefore the five-level Likert scaling was retained.

Bottom Line: The overall AS-20 did not demonstrate unidimensional; however, it was achieved separately in the two Rasch-revised subscales: the psychosocial subscale (11 items) and the function subscale (9 items).The features of good targeting, optimal item infit and outfit, and no notable local dependence were found for each of the subscales.The rating scale was appropriate for the psychosocial subscale but a reduction to four response categories was required for the function subscale.

View Article: PubMed Central - PubMed

Affiliation: School of Nursing, Third Military Medical University, Chongqing, China.

ABSTRACT

Background: The impact of strabismus on visual function, self-image, self-esteem, and social interactions decrease health-related quality of life (HRQoL).The purpose of this study was to evaluate and refine the adult strabismus quality of life questionnaire (AS-20) by using Rasch analysis among Chinese adult patients with strabismus.

Methods: We evaluated the fitness of the AS-20 with Rasch model in Chinese population by assessing unidimensionality, infit and outfit, person and item separation index and reliability, response ordering, targeting and differential item functioning (DIF).

Results: The overall AS-20 did not demonstrate unidimensional; however, it was achieved separately in the two Rasch-revised subscales: the psychosocial subscale (11 items) and the function subscale (9 items). The features of good targeting, optimal item infit and outfit, and no notable local dependence were found for each of the subscales. The rating scale was appropriate for the psychosocial subscale but a reduction to four response categories was required for the function subscale. No significant DIF were revealed for any demographic and clinical factors (e.g., age, gender, and strabismus types).

Conclusion: The AS-20 was demonstrated by Rasch analysis to be a rigorous instrument for measuring health-related quality of life in Chinese strabismus patents if some revisions were made regarding the subscale construct and response options.

No MeSH data available.


Related in: MedlinePlus