Table 2.

Results of the two hierarchical generalized linear models, as implemented in R statistical software, with food habits (Model 1) and hygiene habits (Model 2) as outcomes. Results report odds ratios, confidence intervals, and p values. The factor group was coded 0 for AN_TP1 and AN_TP2 and 1 for HC, therefore the reference group is the patient group and odds ratios of a habit occurring are displayed for HC relative to AN. An odds ratio <1 represents a decreased likelihood in relation to the comparison group, and odds ratio >1 an increased likelihood.

Model 1: Food intake				Model 2: Hygiene
Predictors	Odds Ratios	95 % CI	p	Odds Ratios	95 % CI	p
(Intercept)	36.78	1.40 – 967.93	0.031	20.26	0.79 – 522.35	0.07
Group [HC]	0.23	0.10 – 0.49	<0.001	0.39	0.19 – 0.82	0.013
Time point [TP2]	0.34	0.19 – 0.59	<0.001	0.31	0.19 – 0.51	<0.001
Age	0.92	0.76 – 1.13	0.437	0.89	0.73 – 1.09	0.268
Prompt count	0.99	0.98 – 0.99	<0.001	0.98	0.98 – 0.99	<0.001
Compliance	0.22	0.03 – 1.56	0.129	0.82	0.13 – 5.23	0.834
Random Effects
σ2	3.29			3.29
τ00participant_id:tp	1.5			1.09
τ00participant_id	0.77			0.89
ICC	0.45			0.44
Observations	4069			4110
Marginal R2 / Conditional R2	0.08 / 0.49			0.06 / 0.48