Work outcome in yet undiagnosed patients with non-radiographic axial spondyloarthritis and ankylosing spondylitis; results of a cross-sectional study among patients with chronic low back pain

Table 3 Final results of the ZINB regression model testing associations between demographical and clinical parameters and the WPAI subdomains corrected for age and gender

	Absenteeism		Presenteeism		Work impairment		Activity impairment
Parameter	Count	Logistic	Count	Logistic	Count	Logistic	Count	Logistic
Education level
Intermediate			0.100 (0.195)		-0.051 (0.578)	0.844 (0.004)
High			-0.228 (0.015)		-0.252 (0.017)	0.716 (0.029)
Duration of LBP (years)						0.043 (0.012)
VAS pain		-0.190 (0.008)	0.081 (0.000)	-0.165 (<0.001)	0.079 (<0.001)	-0.324 (<0.001)	0.079 (<0.001)	-0.261 (<0.001)
RMDQ		-0.188 (<0.001)	0.048 (<0.001)	-0.005 (0.744)	0.058 (<0.001)	-0.152 (<0.001)	0.044 (<0.001)	-0.106 (<0.001)

Only the significant regression coefficients are shown, with the p values in parentheses. The logistic part of the ZINB model is generated for the “certain zero” cases, predicting whether or not a patient would be in this group. At the same time the count part of the model is predicting the counts for those patients who are not certain zeros.
ZINB zero inflated negative binomial, WPAI Worker Productivity and Activity Impairment Questionnaire, LBP low back pain, VAS = visual analog scale, RDMQ Roland Morris Disability Questionnaire

ISSN: 1478-6362