Cluster randomised trial of a tailored intervention to improve the management of overweight and obesity in primary care in England

Table 6 Predictive variables

Outcome variable	Predictor	Odds ratio (95 % CI)
Weight management	BMI at baseline^a	1.03 (1.02, 1.05)
	Age^a	1.02 (1.01, 1.03)
BMI or waist circumference measured	Sex (male)^b	0.70 (0.57, 0.87)
BMI or waist circumference measured	Waist circumference [cm]^a	1.01 (1.01, 1.02)
Referral to external weight loss services	Weight at baseline [kg]^a	1.01 (1.01, 1.02)
Internal weight management	Sex (male)^b	1.09 (0.99, 1.20)
	Ethnicity (mixed)^c	0.71 (0.51, 0.99)
	Age^a	1.01 (1.00, 1.02)
Lifestyle assessment	BMI at baseline^a	1.02 (1.01, 1.03)
Weight loss of at least 1 kg	Ethnicity (mixed)^c	0.45 (0.20, 1.00)
	Weight at baseline^a	1.02 (1.01, 1.02)
	BMI at baseline^a	1.03 (1.01, 1.06)
	Age^a	1.02 (1.01, 1.02)
		Increase in outcome (95 % CI)
BMI	Ethnicity (South Asian)^c	−0.47 (−0.97, 0.02)
	BMI at baseline^d	0.74 (0.66, 0.82)
	Age^d	−0.01 (−0.02, −0.00)
Weight	BMI^d	−0.17 (−0.22, −0.12)
	Weight at baseline [kg]^d	0.95 (0.92, 0.97)
	Age^d	−0.07 (−0.08, −0.06)

^aComparison between a binary outcome and continuous predictor. For the proportion of patients offered a weight loss intervention (outcome) and BMI (predictor). The odds ratio of 1.03 implies that an increase in BMI by a unit of 1 leads to a 3 % increase in the odds of receiving a weight management intervention
^bComparison between a binary outcome and categorical predictor. For example, the odds of having a BMI or waist circumference measured is 30 % lower in men compared to women
^cComparison between a continuous outcome and categorical predictor. For example, being of a South Asian ethnicity leads to an increase in BMI of −0.47 (i.e. a decrease of 0.47) in a South Asian when compared to a White European
^dComparison between a continuous outcome and continuous predictor. For example, an increase in BMI at baseline of 1 leads to an increase in BMI at follow-up of 0.74

ISSN: 1748-5908