Development of a predictive model for integrated medical and long-term care resource consumption based on health behaviour: application of healthcare big data of patients with circulatory diseases

Table 1 Correlation coefficient of variables and weighting of predictors

Variable	Correlation coefficient	p value			Weighting
Age	0.42	<0.001
12-month period after enrolment
Health behaviour				Broad adherence
Secondary prevention			1	Secondary prevention (Integrated)	8.18
Health check-ups	0.122	<0.01		*
Item of health check-ups	0.128	<0.001		*
Tertiary prevention
Rehabilitation intensity	0.012	0.43	2	Rehabilitation intensity	0.81
Guidance	0.156	<0.001	3	Guidance	1.04
PDC	0.079	<0.001	4	PDC	5.35
Overlapping outpatient service
Outpatient visits	0.001	0.938	5	Overlapping outpatient visits	3.03
Clinical laboratory and physiological tests	0.049	<0.05	6	Overlapping clinical laboratory and physiological test	5.02
Medical attendance			7	Medical attendance (Integrated)	2.76
Inpatient days	0.52	<0.001		**
Outpatients visits	0.352	<0.001		**
Dispensing	0.354	<0.001		**
Public behaviour
Generic drug rate	0.209	<0.001	8	Generic drug rate	6.58
Long-term care
In-home services, number per year	0.224	<0.001
Community-based services, number per year	0.011	0.466
Facility services, number per year	0.086	<0.001
48-month follow-up period after index 12-month enrolment
Follow-up period	0.442	<0.001
Medical expense	0.938	<0.001
Long-term care expense	0.269	<0.001