Development and validation of a prediction model for home discharge in patients with moderate stroke: The Korean stroke cohort for functioning and rehabilitation study

Figures & data

Figure 1. Patient selection flowchart.

Table 1. Clinical characteristics and univariate analysis of moderate stroke patients.

Factors		Mean±SD or n (%)		OR	95% CI	p-value
Discharge	Home	372	(51)
	Hospitals	360	(49)
Length of hospital stay	Days	32.5	±4.8	1.01	1.00–1.02	0.946
Demographic characteristics
Age		68.7	±13.4	0.96	1.01–1.03	0.601
Gender	Male	415	(57)	1.10	1.05–1.15	0.564
Marital status	Married	616	(84)	1.93	1.54–2.66	0.006*
Education	≥High school	375	(51)	2.34	1.86–2.79	0.128
Smoking	Yes	199	(27)	0.86	0.65–1.11	0.298
Alcohol	Yes	547	(75)	0.78	0.66–0.90	0.772
Family history	Yes	73	(10)	1.06	0.88–1.24	0.557
Comorbidity
Hypertension	Present	484	(66)	0.92	0.65–1.17	0.323
Diabetes mellitus	Present	216	(30)	1.18	0.76–1.60	0.481
Atrial fibrillation	Present	65	(9)	1.72	1.15–2.33	0.452
Hyperlipidemia	Present	129	(18)	0.86	0.36–1.36	0.921
BMI	≥30 kg/m^2	65	(9)	0.98	0.93–1.03	0.831
mCCI	Low (0–1)	414	(57)	4.80	3.63–6.35	0.003*
	High (2–37)	318	(43)	1.00
Stroke features
Stroke type	Ischemic	540	(74)	2.83	1.45–4.30	0.090*
	Hemorrhagic	192	(26)
Lesion side	Right	412	(56)	3.32	1.20–5.33	0.074*
Day 7 NIHSS		14.2	±5.8	1.11	1.02–1.20	0.195
Functional status at discharge
Total FIM		86.2	±23.8	1.05	1.03–1.07	0.220
Motor FIM	Low (13–38)	197	(27)	1.00
	Medium (39–50)	245	(33)	2.39	1.06–3.87
	High (51–91)	290	(40)	5.05.	2.03–8.21	0.076*
Cognitive FIM	Low (5–20)	160	(22)	1.00
	Medium (21–29)	191	(26)	4.80	2.64–7.48
	High (30–35)	381	(52)	15.28	10.13–20.16	<0.001*
NIHSS		9.1	±5.6	0.89	0.75–1.05	0.393
K-MMSE	Normal (25–30)	389	(53)	1.36	1.05–1.67	0.411
	Abnormal (0–24)	343	(47)	1.00
FMA	Normal (80–100)	179	(24)	1.44	1.32–1.56	0.263
	Abnormal (0–79)	553	(76)	1.00
ASHA-NOMS	Normal (7)	480	(66)	2.34	3.05–8.10	0.075*
	Abnormal (0–6)	252	(34)	1.00
K-FAST	Normal (26–30)	422	(58)	2.78	1.55–4.13	0.180
	Abnormal (0–25)	310	(42)	1.00
FAC	Independent (4–5)	251	(34)	11.36	8.08–14.53	<0.001*
	Dependent (0–3)	481	(66)	1.00
SGDS-K	Normal (0–9)	351	(48)	2.35	1.90–2.78	0.133
	Abnormal (10–15)	381	(52)	1.00

Abbreviations: BMI, body mass index; mCCI, modified Charlson Comorbidity Index; FIM, Functional Independence Measure; NIHSS, National Institutes of Health Stroke Scale; K-MMSE, Korean version of the Mini-Mental State Examination; FMA, Fugl-Meyer Assessment; ASHA-NOMS, American Speech-Language-Hearing Association National Outcome Measurement System swallowing scale; K-FAST, Korean version of the Frenchay Aphasia Screening Test; FAC, Functional Ambulation Categories; SGDS-K, Korean version of the Geriatric Depression Scale-Short Form; OR, odds ratio.

*p < 0.10

Table 2. Multivariate model for home discharge and weighted scoring system for moderate stroke patients.

Variables		B (SE)†	p-value	OR	95% CI	Score
Intercept		−2.53 (0.36)	<0.001*
Marital status	Single	0				0
	Married	0.61 (0.27)	0.038*	1.84	1.34–2.34	1
mCI	High (2–37)	0				0
	Low (0–1)	1.33 (0.31)	0.016*	3.21	1.73–4.76	2
Cognitive FIM	Low (5–20)	0				0
	Medium (21–29)	1.58 (0.33)	<0.001*	3.82	2.48–5.43	3
	High (30–35)	3.72 (0.36)	<0.001*	8.43	5.13–12.34	6
FAC	Dependent (0–3)	0				0
	Independent (4–5)	2.35 (0.30)	<0.001*	6.80	4.48–10.23	4

Abbreviations: SE, standard error; mCCI, modified Charlson Comorbidity Index; FIM, Functional Independence Measure; FAC, Functional Ambulation Categories.

*p < 0.05

†Coefficient B represents the weight of each parameter.

Figure 2. The receiver operating characteristic curve after applying the model in this study to (A) development data set (AUC = 0.87; 95% CI = 0.84–0.90) and (B) the external data set (AUC = 0.86; 95% CI = 0.80–0.92).