Table 2.

Logistic regression analysis results for significant characteristics between users and non-users of CAM and prayer-for-health

	Multivariate logistic regression model of
	CAM users vs. non-CAM users† (Included prayer-for-health)	CAM users vs. non-CAM users‡ (Excluded prayer-for-health)	Users vs. non-users§ (Prayer-for-health only)
Characteristics	OR (95% CI)	OR (95% CI)	OR (95% CI)
Ethnicity
Malay	-	-	Reference
Chinese			0.31 (0.20–0.49)**
Indian			0.81 (0.46–1.42)
Others			0.57 (0.15–2.18)
Education level
Primary school/lower	Reference	Reference	-
Secondary school	1.44 (0.90–2.24)	1.28 (0.82–1.99)
Tertiary	2.33 (1.22–4.47)*	2.11 (1.15–3.89)*
Household income/month
≤RM 3000 (≈944 USD)	Reference	Reference	Reference
>RM 3000 (≈944 USD)	3.41 (1.92–6.03)**	2.32 (1.40–3.84)**	2.55 (1.66–3.93)**
Staging of disease
Early	Reference	Reference	Reference
Advanced	1.86 (1.23–2.82)**	1.75 (1.18–2.59)*	2.10 (1.39–3.13)**
Chemotherapy adherence
Postponed	Reference	Reference	-
On schedule	0.18 (0.06–0.51)**	0.24 (0.10–0.58)**

†Logistic regression model: Hosmer and Lemeshow test, χ² = 3.47 (5), p = 0.63.

‡Logistic regression model: Hosmer and Lemeshow test, χ² = 8.46 (6), p = 0.21.

§Logistic regression model: Hosmer and Lemeshow test, χ² = 11.43 (6), p = 0.08.

**p < 0.001.

*p < 0.05.