. 2023 Oct 25;63(8):2152–2161. doi: 10.1093/rheumatology/kead566

Table 3.

Multivariable logistic regression

(A) Factors associated with PsA diagnosis
	Analysis including ‘drop-out’ patients^a			Analysis NOT including ‘drop-out’ patients^b
	OR	95% CI	P	OR	95% CI	P
Gender	2.023	0.999–4.091	0.050	2.065	1.009–4.228	0.047
Female vs male	2.023	0.999–4.091	0.050	2.065	1.009–4.228	0.047
Previous b/tsDMARDs use	0.264	0.104–0.674	0.005	0.256	0.100–0.659	0.005
Present vs absent	0.264	0.104–0.674	0.005	0.256	0.100–0.659	0.005

(B) Factors associated with axPsA final diagnosis
	OR	95% CI	P\|	OR	95% CI	P
Gender	2.030	0.880–4.682	0.097	2.034	0.875–4.728	0.099
Female vs male	2.030	0.880–4.682	0.097	2.034	0.875–4.728	0.099
Onychopathy	2.966	1.139–7.723	0.026	3.024	1.156–7.912	0.024
Present vs absent	2.966	1.139–7.723	0.026	3.024	1.156–7.912	0.024
Pso duration (years)	0.971	0.943–1.002	0.068	0.973	0.943–1.003	0.078

(C) Factors associated with pPsA final diagnosis
	OR	95% CI	P	OR	95% CI	P
Gender	2.236	1.059–4.722	0.035	2.286	1.072–4.876	0.035
Female vs male	2.236	1.059–4.722	0.035	2.286	1.072–4.876	0.035
Previous b/tsDMARDs use	0.349	0.135–0.901	0.030	0.342	0.132–0.891	0.028
Present vs absent	0.349	0.135–0.901	0.030	0.342	0.132–0.891	0.028

A backward stepwise logistic regression was conducted to identify potential predictors of the outcomes from a poll of candidate variables, including gender, age, BMI, active smoking habit, PASI, duration of skin psoriasis, presence of nail disease and previous biologic use. At each step, variables were removed based on P-values, with a threshold of 0.01 used to determine the final set of variables in the model. The table displays the results of two logistic regression analyses:

one including eligible patients who declined the rheumatological evaluation (on the left), and

one excluding this group (on the right). In the first analysis, the ‘drop-out’ patients were classified as having psoriasis without a diagnosis of PsA. Results are shown as follows: (A) probability of being diagnosed with PsA, (B) probability of being diagnosed with axPsA and (C) probability of being diagnosed with pPsA, for the patients enrolled in the ATTRACT (Axial psoriaTic arThritis scReening AnCona iTaly) study. The P-value was considered significant if <0.05, and significant P-values are shown in bold. Statistical analyses were conducted using ‘Stata’ software. OR: odds ratio; axPsA: axial PsA; pPsA: peripheral PsA; Pso: psoriasis; b/tsDMARDs: biologic and/or targeted synthetic DMARDs.