Purchase this article with an account.
Sashia Bak-Nielsen, Therese Koops Groenborg, Anders Ivarsen, Jesper Hjortdal; Keratoconus Outcome Research Questionnaire – a validation study of the Danish version. Invest. Ophthalmol. Vis. Sci. 2018;59(9):4395.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Reduced vision-related quality of life (VR-QoL) has been reported among Keratoconus (KC) patients. However, the validity of the patient reported outcome (PRO) instruments has been questioned, as they are neither disease-specific nor satisfy modern scale validation. The Keratoconus Outcome Research Questionnaire (KORQ) is the first instrument developed specifically to measure two domains of VR-QoL in KC patients. The KORQ consist of 2 scales, the activity limitation scale (AL-S) (18 items) and the symptoms scale (S-S) (11 items). The purpose of this study was to translate the KORQ into Danish language and validate it in a population of Danish KC patients.
The KORQ was translated using standardised translation techniques. 191 KC patients completed the KORQ between December 2016 and August 2017. Data were analysed using a polytomous rating scale Rasch model.
Participants had a mean age of 34.5 (± 12.1) years and 75.4 % were male. KC severity (Pentacam grading): KC 1: 7.1 % KC 2: 35.5 % KC 3: 52.5 % KC 4: 4.9%.Inter-item correlation range: 0.21-0.83 (AL-S) and 0.00-0.73 (S-S), though with the majority above 0.30. Two items (S4 and S5) showed low inter-item correlation with the remaining items in the S-S.The Rasch analysis showed ordered threshold in all items, with a small threshold span for answer category 3 (fig. 1, 2). The overall item discrimination was 1.56 (95%CI: 1.37-1.76) (AL-S) and 0.93 (95%CI: 0.80 – 1.07) (S-S). Person separation index was 0.93 (AS-L) and 0.83 (S-S). All items in the AL-S showed perfect fit, with item mean square between 0.69 and 1.25. For the S-S, all items except S4 had item mean square between 0.5 and 1.5.
The Danish version of the KORQ showed good inter-item correlation except for 2 items. The Rasch analysis showed ordered item thresholds. The item discrimination and person separation index were considerable higher for the AL-S compared to the S-S. This indicates that a reconstruction of the S-S may be necessary for the Danish version of the KORQ to be psychometrically robust.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.
Fig1: Threshold map of items in serial order, the most likely answer in each item given a certain level of VR-QoL. Activity limitation scale: A1-A18, Symptoms Scale: S1-S11.
Fig 2: Category Characteristic Curves for item A1. The probability of answering category 1-4 given a certain level of the underlying trait (VR-QoL).
This PDF is available to Subscribers Only