Table 3.

Multivariable analysis of depression and anxiety scores

	Depression score		Anxiety score
	Coefficient (95% CI)	P value	Coefficient (95% CI)	P value
OSDI (per 10 unit increase)	12% (7% to 16%)	<0.001	12% (7% to 16%)	<0.001
Employment		0.040		N.S.
Full/part/self	–	–	–	–
Student/unemployed	48% (11% to 96%)	0.007	–	–
House-wife/husband	28% (−6% to 76%)	0.121	–	–
Retired	21% (−3% to 52%)	0.093	–	–
Eye drops per day		0.300		0.009
None or as required	–	–	–	–
1–4	20% (−7% to 54%)	0.153	36% (5% to 77%)	0.022
5+	7% (−19% to 41%)	0.651	57% (17% to 110%)	0.003

Results are from general linear models, with the log-transformed depression/anxiety scores as the dependent variables. A backward stepwise approach was used to select variables, with the following considered for inclusion: age, gender, ethnicity, education, employment, disease subgroup (OSD/Uveitis), OSDI, steroid use and eye drop frequency. The coefficients were then anti-logged and represent the percentage increase in the anxiety/depression score for the stated category relative to the reference, or for a 10-unit increase in OSDI. Bold p values are significant at p<0.05. N.S.=not selected by the stepwise procedure for inclusion in the final model.

OSD, ocular surface disease.