Development of a new health-related quality of life measure for people with diabetes who experience hypoglycaemia: the Hypo-RESOLVE QoL

Carlton, Jill; Powell, Philip A.; Broadley, Melanie; de Galan, Bastiaan E.; Heller, Simon; Comins, Jonathan; Rosilio, Myriam; Pouwer, Frans; Gall, Mari-Anne; Child, Christopher J.; McCrimmon, Rory J.; Rowen, Donna

doi:10.1007/s00125-024-06182-9

Table 2 Psychometric analyses used to help inform item selection (Stage 3)

From: Development of a new health-related quality of life measure for people with diabetes who experience hypoglycaemia: the Hypo-RESOLVE QoL

Analysis	Benchmarking criteria used to determine acceptable performance
Descriptives
Floor and ceiling effects	>5% and <40% of responses in the floor/ceiling response category
Non-applicable/missing responses	<20% of responses are non-applicable, <5% of responses are missing (paper surveys only)
Mokken scale analysis^a
Coefficients of homogeneity (H)	H≥0.30 [34]
Local independence (conditional association)	Item does not significantly violate local independence assumption [34]
Monotonicity	Item does not significantly violate monotonicity (critical value ≥80) [35]
IRT partial credit model^b
Item fit (infit and outfit)	Mean squares between 0.6 and 1.4 [36]
Item category thresholds	Item category threshold parameters are ordered [37]
Differential item functioning (DIF)^c	Non-significant DIF (α<0.01) in logistic regression with IRT θ estimates as the conditioning variable, and McFadden R²<0.02 [38]
Classical test theory
Confirmatory factor analysis, with a one-factor baseline and three-factor hypothesised model (physical, psychological, social)^d	CFI ≥0.9 [29]; RMSEA <0.08 [39]
Cronbach’s α^e	Cronbach’s α ≥0.7 [40]
Item correlations
Polychoric intercorrelations between items^e	Intercorrelations between items <0.7, indicating non-redundancy [41]
Spearman correlations between items and the hypoglycaemia-related QoL VAS (4 weeks)^e	Non-trivial correlation (r_s≥0.2) [42]
Spearman correlations between items and HypoA-Q question 1 (i.e. number of hypos in the previous week)^e	Non-trivial correlation (r_s≥0.2) [42]

R package dplyr was used to help clean and recode data prior to analysis
^aPsychometric analyses were conducted using R package mokken
^bPsychometric analyses were conducted using R package eRm
^cPsychometric analyses were conducted using R package lordif
^dPsychometric analyses were conducted using R package lavaan
^ePsychometric analyses were conducted using R package psych

Back to article page