Table 3. Associations of the types of student financial support and the total amount of student loan debt with severe psychological distress^a from Poisson regression models with a robust error variance stratified by current student status after imputation.

			Model 1b		Model 2c
			PR	95% CI	PR	95% CI
Among graduates and dropouts			(n = 3,170)		(n = 3,170)
Types of student financial support		None	Reference		Reference
		Student loans	1.22	1.01, 1.47	1.26	1.04, 1.53
		Scholarships	0.84	0.41, 1.70	0.87	0.43, 1.77
		Both types	1.10	0.55, 2.22	1.18	0.59, 2.39
	Among graduates and dropouts with only student loan and none		(n = 3,053)		(n = 3,053)
	Total amounts of student loan debt	None	Reference		Reference
		<2.0 million yen	1.02	0.76, 1.39	1.04	0.77, 1.41
		2.0 to 4.0 million yen	1.21	0.93, 1.58	1.26	0.96, 1.65
		≥4.0 million yen	1.38	0.98, 1.94	1.44	1.02, 2.03
Current university students			(n = 979)		(n = 979)
Types of student financial support		None	Reference		Reference
		Student loans	0.97	0.65, 1.44	0.91	0.60, 1.37
		Scholarships	0.66	0.17, 2.58	0.72	0.18, 2.88
		Both types	1.09	0.35, 3.42	1.12	0.36, 3.52

CI, confidence interval; PR, prevalence ratio.

^aSevere psychological distress was assessed by the Kessler Psychological Distress Scale by a cut-off point at 12/13.

^bModel 1: Age and sex were adjusted.

^cModel 2: Model 1 + educational attainment, sources of the enrolled university, father’s educational attainment, mother’s educational attainment, and current parents’ annual household income were adjusted.

The single imputation was conducted using types of student financial support, amounts of student loan debt, current student status, age, sex, educational attainment, sources of the enrolled university, father’s educational attainment, mother’s educational attainment, current parents’ annual household income, and the K6 score by the the k-nearest neighbor algorithm from the R package named “DMwR.”