Table 3.

Personal walking habits of participants by community.

	N	Prev. (%)	Model 1  unadjusted		Model 2 adjusted for personal characteristics2		Model 3 adjusted for personal and distance to amenities3		Model 4 adjusted for personal, distance, and community SES4
			OR1	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
(A) Walking ≥5 days/week
Boston: Hyde Park	36	16.7	Ref.		Ref.		Ref.		Ref.
Boston: Mattapan	24	25.0	1.67	0.47–5.96	2.02	0.54–7.62	2.82	0.74–10.79	2.36	0.62–8.93
Boston: West Roxbury	71	25.4	1.70	0.61–4.74	1.96	0.67–5.68	2.57	0.87–7.57	2.40	0.81–7.17
Newton-Central	27	25.9	1.75	0.51–5.98	1.92	0.52–7.06	2.45	0.66–9.06	1.88	0.50–6.99
Boston: North/South Dorchester	60	26.7	1.82	0.64–5.18	2.39	0.80–7.12	2.25	0.74–6.80	2.08	0.69–6.28
Brookline-South	51	31.4	2.29	0.79–6.58	1.70	0.57–5.10	1.41	0.46–4.26	1.03	0.33–3.26
Newton-West	73	31.5	2.30	0.84–6.29	1.87	0.66–5.29	3.29	1.15–9.45∗	2.48	0.85–7.23
Dedham, Needham	40	32.5	2.41	0.80–7.22	1.95	0.62–6.08	2.90	0.92–9.16	1.63	0.52–5.16
Milton	47	34.0	2.58	0.89–7.48	2.68	0.89–8.11	3.50	1.14–10.76∗	2.72	0.89–8.31
Boston: Roslindale	37	35.1	2.71	0.90–8.19	3.33	1.04–10.66∗	3.55	1.09–11.56∗	3.41	1.05–11.10∗
Boston: Roxbury	59	35.6	2.76	0.99–7.71	3.31	1.13–9.69∗	2.82	0.96–8.31	2.23	0.74–6.72
Boston: Allston/Brighton	38	36.8	2.92	0.97–8.73	2.75	0.88–8.56	1.74	0.55–5.49	1.44	0.43–4.78
Newton-East	47	38.3	3.10	1.08–8.92∗	2.27	0.76–6.82	2.82	0.93–8.54	1.82	0.60–5.57
Brookline-North	51	47.1	4.44	1.58–12.51∗∗	3.89	1.32–11.50∗	2.29	0.76–6.85	1.64	0.49–5.44
Boston: Jamaica Plain	57	47.4	4.50	1.62–12.47∗∗	4.95	1.70–14.43∗∗	3.37	1.15–9.89∗	2.66	0.86–8.25
Boston: Downtown	27	48.1	4.64	1.46–14.76∗∗	6.86	2.01–23.38∗∗	3.24	0.95–11.09	2.22	0.55–8.99
	745
Overall p value for community differences				p = 0.10		p = 0.07		p = 0.56		p = 1.0
Area under ROC curve				0.60		0.73		0.74		0.75
p value for change in area under ROC compared to previous model				NA		p < 0.001		p = 0.004		p = 0.64
(B) Walking for recreation ≥1 time/week
Boston: Hyde Park	36	41.7	1.46	0.64–3.35	1.52	0.65–3.58	1.10	0.46–2.63	0.95	0.39–2.27
Boston: Mattapan	24	41.7	1.46	0.56–3.78	1.70	0.62–4.66	1.36	0.49–3.76	1.14	0.41–3.16
Boston: West Roxbury	70	32.9	Ref.		Ref.		Ref.		Ref.
Newton-Central	26	38.5	1.28	0.50–3.25	1.14	0.43–3.00	1.04	0.39–2.74	0.79	0.30–2.11
Boston: North/South Dorchester	58	41.4	1.44	0.70–2.97	1.67	0.79–3.54	1.16	0.55–2.48	0.97	0.46–2.08
Brookline-South	50	46.0	1.74	0.82–3.67	1.44	0.66–3.12	0.88	0.40–1.95	0.86	0.39–1.92
Newton-West	73	38.4	1.27	0.64–2.53	1.20	0.59–2.45	1.14	0.55–2.33	1.06	0.52–2.18
Dedham, Needham	40	40.0	1.36	0.61–3.05	1.41	0.60–3.27	1.23	0.53–2.87	1.02	0.44–2.40
Milton	47	34.0	1.05	0.48–2.31	1.01	0.45–2.26	0.78	0.34–1.77	0.72	0.32–1.65
Boston: Roslindale	37	43.2	1.56	0.69–3.53	1.68	0.72–3.92	1.25	0.53–2.94	1.04	0.44–2.46
Boston: Roxbury	59	39.0	1.31	0.63–2.69	1.30	0.62–2.75	0.97	0.46–2.07	0.86	0.40–1.84
Boston: Allston/Brighton	38	47.4	1.84	0.82–4.13	1.59	0.68–3.69	0.96	0.41–2.28	0.99	0.42–2.33
Newton-East	47	53.2	2.32	1.09–4.96∗	1.72	0.79–3.77	1.33	0.59–2.96	1.00	0.44–2.27
Brookline-North	51	49.0	1.96	0.94–4.13	1.87	0.86–4.05	1.09	0.49–2.42	0.96	0.43–2.14
Boston: Jamaica Plain	56	44.6	1.65	0.80–3.40	1.63	0.77–3.46	0.99	0.46–2.14	0.93	0.43–2.01
Boston: Downtown	27	48.1	1.90	0.77–4.69	1.83	0.71–4.72	1.01	0.39–2.66	0.92	0.35–2.43
	739
Overall p value for community differences				p = 0.86		p = 0.97		p = 1.0		p = 1.0
Area under ROC curve				0.56		0.66		0.68		0.68
p value for change in area under ROC compared to previous model				NA		p < 0.001		p = 0.007		p = 0.92

^∗ p < 0.05.

^∗∗ p < 0.01.

¹Odds ratio estimates for the specified walking habit relative to a member of the referent community, obtained by logistic regression.

²Logistic regression model adjusted for a composite score predicted from the following personal characteristics: age, sex, self-rated health, bodily pain, alcohol consumption, education level, body mass index, short physical performance battery, falls efficacy, ADL (activities of daily life ability), race, foot pain, balance, number of comorbidities, gait speed, peripheral neuropathy, flights of stairs in the home, Mini-Mental State Examination Score, number of medications, illness causing reduced activity in past year, and strength to rise from a chair.

³Logistic regression model adjusted for a composite score predicted from the personal characteristics as listed in model 2 and for distances from the participant's block centroid to the following amenities: the nearest bus stop, subway station, hospital, shopping center, post office, public park, food store, town hall, and library.

⁴Logistic regression model adjusted for a composite score predicted from the personal and distances-to-amenities characteristics as listed in model 3 and for the following socioeconomic characteristics of the participant's community: median household income, percent below federal poverty level, percent of adults unemployed, percent of college graduates, percent of housing units owner-occupied, percent of housing units vacant, and percent of minority (non-White).