Quantifying the relationship between gardening and health and well-being in the UK: a survey during the covid-19 pandemic

Table 1 Odds Ratios (OR) for ‘not good health’ as compared to ‘good health’ ^{a, b, c, d}

Variable (reference category)	B (SE)	OR (95% CI)	p value
Constant	-2.04 (0.63)	0.13 (0.03–0.41)	< 0.01
Gender (Female)
Male	-0.46 (0.67)	0.63 (0.16–2.24)	0.49
Age (18–34)
35–54	0.34 (0.73)	1.40 (0.33–6.00)	0.65
55+	-0.27 (0.83)	0.76 (0.15–3.96)	0.74
Obesity (Not obese)
Obese	2.08 (0.61)	7.98 (2.49–27.77)	< 0.001
Long-term conditions (No)
Yes	2.92 (0.67)	18.57 (5.52–79.10)	< 0.001
Physical activity level (Low)
Moderate	-1.19 (0.69)	0.30 (0.07–1.08)	0.08
High	-0.70 (0.94)	0.50 (0.06–2.73)	0.46
Food growing level (No food grown)
1 (very little F&V)	-1.08 (0.91)	0.34 (0.05–1.87)	0.23
2	-1.53 (0.91)	0.22 (0.03–1.17)	0.09
3	-1.99 (0.83)	0.14 (0.02–0.64)	< 0.05
4	-3.27 (1.37)	0.04 (0.00–0.41)	< 0.05
5 (nearly self-sufficient in F&V)	-2.07 (1.01)	0.13 (0.01–0.83)	< 0.05

^a ‘Not good health’ refers to responses of ‘poor’ or ‘fair’ on the SRH, ‘good health’ includes responses of ‘good’, ‘very good’ or ‘excellent’
^b Predictors and regression coefficients in the table are derived from the best fit model for the outcome based on the Bayesian Information Criterion (BIC). Other explanatory variables tested include neighbourhood deprivation, household income, household composition, caring responsibilities, higher education, smoking status, alcohol consumption, F&V intake, time spent gardening, and having allotment, but these were dropped in the process of improving model fit
^c Model R² = 0.39 (Hosmer Lemeshow), 0.30 (Cox and Snell), 0.50 (Nagelkerke); χ² [12] = 66.45
^d Figures in bold are statistically significant at the 5% level (p < 0.05)

ISSN: 1471-2458