Table 4.

Logistic regression of predictors of medication knowledge deficiencies.

	Univariate Analysis a,b				Multiple Logistic Regression e (n = 390)
	Yes c	No d	p	OR [95% CI]	OR [95% CI]
Number of prescribed medicines at discharge (mean [SD])	n = 337	n = 57			1.17 [1.05–1.30]
	9.5 [3.45]	7.9 [2.58]	0.001	1.18 [1.07–1.31]
Geriatric risk profile (score 0–6) (mean [SD])	n = 330	n = 53			/
	2.7 [0.96]	2.4 [0.87]	0.039	1.41 [1.02–1.95]
In-hospital medication management	n = 335	n = 56
By nurses (%)	76.4	58.9	0.007	2.26 [1.25–4.07]	/
By patients (full/partial) (%)	23.6	41.1		Ref.
Patient was … in hospital to manage medication at home	n = 334	n = 56
Not/insufficiently prepared (%)	56.6	37.5	0.009	2.17 [1.21–3.89]	2.02 [1.12–3.64]
Sufficiently prepared (%)	43.4	62.5		Ref.

Note: CI = confidence interval; ref = reference, OR = odds ratio; / = these variables were not included in the model. ^a All patient and medication management characteristics were studied in univariate analysis. Only significant variables were presented in this table. ^b Deviating sample size due to missing data (missing completely at random) [29]. ^c Patients with medication knowledge deficiencies after discharge. ^d Patients without medication knowledge deficiencies after discharge. e Multiple logistic regression analysis: Nagelkerke R²: 0.076, p < 0.001. Colored background: This ensures uniformity as deviating n-values have a colored background.