Developing a short-form version of the HIV Disability Questionnaire (SF-HDQ) for use in clinical practice: a Rasch analysis

Table 7 Social Inclusion Domain (7 items) - Overview of final SF-HDQ item level and domain model fit statistics

Item #	Final SF-HDQ items	Item level statistics
Social domain		Location	SE	Residual	ChiSq χ²	p value
58	Find hard to meet the needs of those I care for	0.808	0.040	− 0.302	11.066	0.271
62	Feel work performance limited (0 = not at all or not applicable)	0.046	0.035	0.455	5.823	0.757
63	Struggle to maintain safe and stable housing - DIF(country)^a	0.570	0.060	− 0.297	22.579	0.007
65	Find it hard to ask others for help when go through an episode of illness - DIF(country)	− 0.310	0.034	− 0.442	7.370	0.599
67	Find it hard to start new intimate, sexual relationships living with HIV (0 = not at all or not applicable)^a	− 0.609	0.058	2.426	8.228	0.511
68	Tend to isolate myself from others because I am HIV positive	− 0.172	0.034	− 3.142	24.214	0.003*
69	Find it hard to take part in leisure or recreational things because can’t afford it - DIF(country)	− 0.333	0.033	0.134	6.398	0.700
Final model: social domain—7 items (raw score range: 0–24)
Mean		− 0.724
Standard deviation		1.002
Sample size		943
Chi-square statistic (df) p value		86.6772 df: 54; p = 0.025679
Cronbach’s alpha		0.79355
Person Separation Index		0.74478
Unidimensionality t-test (% significant)		1.27%
Differential Item Functioniong (DIF)		Significant DIF for country (defined as F value significant and > 1.0 logit difference between groups)—3 items (HDQ63; HDQ65; HDQ69)^b

^*Residuals: HDQ68 (isolate self)—significant residual > -2.5 (p < 0.004; Bonferroni: 0.05/11). This item may be referring to a ‘strategy’ (isolating self) that may lead to challenges to social inclusion rather than social inclusion itself. However we retained due to clinical importance, and because if we remove this item from the model, it worsens fit (PSI = 0.676; model not shown)
Rescoring: ^aRescored to 3 response categories (0–2)—HDQ63 and HDQ67 to achieve ordered thresholds
HDQ61: deleted because captured in HDQ62—which describes limitations rather than prevention (all or nothing)
^bDifferential Item Functioning (DIF): Significant DIF for HDQ63, HDQ65, and HDQ69. Items retained given clinical importance and expected cultural differences related to disability between samples of Canadian and Irish participants

ISSN: 1477-7525