Table 3.

The association between polypharmacy and healthcare cost (KRW)

		Regression estimate for polypharmacy (standard error)
	Multivariable regression				Two-part model
Reference Non-polypharmacy Comparison Polypharmacy	Total cost1	Outpatient costs1	Drug costs 1		Inpatient costs2
Total (n = 3,297)	872,018 (153,868) *	324,879 (51,512) *	536,721 (20,589) *		1,282,226 (569,015) *
65–79 (n = 2,861)	847,661 (136,103) *	359,304 (54,547) *	534,199 (21,104) *		1,340,837 (447,457) *
≥ 80 (n = 436)	825,784 (771,406)	138,873 (156,864)	529,577 (74,149) *		1,056,337 (3,288,392)
Male (n = 1,488)	1,034,734 (278,886) *	390,089 (91,796) *	634,259 (32,586) *		1,213,485 (1,008,922)
Female (n = 1,809)	703,429 (160,331) *	270,219 (55,606) *	451,030 (26,085) *		1,257,406 (591,987) *
Living alone (n = 707)	1,080,115 (314,014) *	355,408 (133,679) *	494,303 (43,979) *		2,851,996 (897,463) *
Not living alone (n = 2,590)	819,306 (176,264) *	327,679 (54,811) *	548,063 (23,362) *		857,487 (693,495) *

Notes. The two-part model for inpatient costs included 748 people who had hospitalization experience, of whom 641 were 65–79 years old, 107 were ≥ 80 years old, 354 were male, 394 were female, 177 lived alone, and 571 did not live alone. Regression estimates for the two-part model was adjusted for the probability of hospitalization. All models controlled for age (years), sex, residential areas, income quartile (household), health insurance type, having private medical insurance, health screening in 2 years, cancer screening in 2 years, education level, Charlson comorbidity index, disability, BMI, living alone, walking days per week, having ever smoked, non-excessive drinking, and calendar year. Standard errors are in brackets. *p < 0.05