Claim at v1.2.0
mob-active-commute-obesity
a7b2be880f03d9d1a8fbf29744852cc1ff3ccb071345bf446cc6ef01318d08f8
Data
{
"has_outcome": true,
"kind": "confirmation",
"lineage": {
"audit": "docs/lineage_audit.md; data/derived/lineage_audit.json",
"class": "none",
"class_meaning": "the audit found no city-data source fetched before registration",
"classes": [],
"early_sources": []
},
"outcome": "mob-active-commute-obesity",
"prior": 0.5,
"registered_at": "2026-11-12T15:25:31-0500",
"registrant": "mobility agent (Claude Opus 5.5), for J. Vernon",
"registration": {
"alternatives": [
"Income, education and race/ethnicity composition (poverty is the named confounder; others are not measured here)",
"Density and Manhattan: walking to work is common where jobs are close, which is also where incomes are high",
"Reverse causation and selection: people with lower BMI choose to walk or cycle, or choose walkable neighbourhoods",
"Commute mode covers workers only; obesity covers all adults"
],
"graph_evidence": [
{
"concepts": [
"physical activity",
"obesity"
],
"direction": "+",
"n_findings": 82,
"n_studies": 58,
"pubmed_ids": [
11985949,
15229327,
20368941,
21605360,
23308135,
23844958,
23958406,
24282316,
24406093,
24918414,
26477996,
27924618,
28404503,
30688830,
30921554,
34274952,
34556954,
38207152
],
"relationship_id": "dPeET-pgVN|cj7G775u_Q",
"summary": "21 ratio findings; 11 above 1, 10 below; 6 with CI above 1, 7 with CI below 1; median ratio 1.2 (mixed increments and designs)"
},
{
"concepts": [
"walking",
"obesity"
],
"direction": "+",
"n_findings": 3,
"n_studies": 2,
"pubmed_ids": [
33225582,
37056656
],
"relationship_id": "_2dhcsFISl|cj7G775u_Q",
"summary": "3 ratio findings; 2 above 1, 1 below; 2 with CI above 1, 1 with CI below 1; median ratio 1.96 (mixed increments and designs)"
},
{
"note": "The graph's physical activity -> obesity findings are split in sign (the ratio findings mix 'activity predicts lower obesity' with 'inactivity predicts higher obesity', so the sign of a ratio depends on the coding of the exposure). There is no concept for cycling or active commuting; variable search finds a few commuting variables (e.g. 'active commuting to/from school', 'being a non-active commuter')."
}
],
"id": "mob-active-commute-obesity",
"kind": "confirmation",
"literature": [
{
"citation": "Flint E, Cummins S, Sacker A (2014). Associations between active commuting, body fat, and body mass index: population based, cross sectional study in the United Kingdom. BMJ 349:g4887.",
"doi": "10.1136/bmj.g4887",
"finding": "Men commuting by active or public modes had BMI about 1 kg/m2 lower than car commuters; women about 0.7 kg/m2."
},
{
"citation": "Celis-Morales CA et al. (2017). Association between active commuting and incident cardiovascular disease, cancer, and mortality. BMJ 357:j1456.",
"doi": "10.1136/bmj.j1456",
"finding": "Cycle commuting associated with lower CVD incidence and mortality (UK Biobank)."
}
],
"plan": "Data: EHDP (github.com/nychealth/EHDP-data, commit 547482b9df589006543e13e8cdf6546e00ed9998): indicator 2415 commuting modes, measures 1304 (bicycle %) + 1307 (walked %), UHF42, ACS 2015-19; indicator 2063 adult obesity %, measure 506, UHF42, mean of 2017, 2018 and 2019 (CHS, self-reported BMI); indicator 103 poverty %, measure 221, UHF42, 2015-19. Model: OLS of obesity % on active-commute % (per 10 points), then adding poverty %; HC3 standard errors; sensitivity adding borough fixed effects and excluding Manhattan. Figure: scatter of UHF neighbourhoods coloured by poverty with the fitted adjusted slope and the slope implied by Flint et al. (about -1 kg/m2 BMI for active commuters, translated to obesity prevalence only qualitatively).",
"prediction": {
"direction": "-",
"magnitude": "Unadjusted slope: -2 to -6 percentage points of adult obesity per 10 points of walk-or-bike commute share. Poverty-adjusted: -0.5 to -4 points per 10 points.",
"refute_if": "Poverty-adjusted slope at or above 0 with an interval that excludes -0.5.",
"support_if": "Poverty-adjusted slope below 0 with a 95% interval below 0."
},
"prior": 0.5,
"registered_at": "2026-11-12T15:25:31-0500",
"registrant": "mobility agent (Claude Opus 5.5), for J. Vernon",
"sha256": "ae99f6c6638ffddcd7df7f9576b3f372e9359d98dd86b546a4c396e9a1ba80ae",
"statement": "Across New York City's UHF42 neighbourhoods, a higher share of workers who walk or cycle to work goes with a lower share of adults with obesity, after adjusting for neighbourhood poverty."
},
"registration_commit": {
"commit": "f99a06e7f3fcda921080138c6d51ab92789e947e",
"committed_at": "2026-11-12T15:25:31-05:00"
},
"registration_file_sha256": "6951d38110f1fb40623681207f01a9e31c612b199daad1e1c9f14cbefe5671fd",
"registration_path": "claims/registered/mob-active-commute-obesity.json",
"registration_sha256": "ae99f6c6638ffddcd7df7f9576b3f372e9359d98dd86b546a4c396e9a1ba80ae",
"registration_text": "{\n \"id\": \"mob-active-commute-obesity\",\n \"kind\": \"confirmation\",\n \"statement\": \"Across New York City's UHF42 neighbourhoods, a higher share of workers who walk or cycle to work goes with a lower share of adults with obesity, after adjusting for neighbourhood poverty.\",\n \"graph_evidence\": [\n {\n \"relationship_id\": \"dPeET-pgVN|cj7G775u_Q\",\n \"concepts\": [\n \"physical activity\",\n \"obesity\"\n ],\n \"n_findings\": 82,\n \"n_studies\": 58,\n \"direction\": \"+\",\n \"summary\": \"21 ratio findings; 11 above 1, 10 below; 6 with CI above 1, 7 with CI below 1; median ratio 1.2 (mixed increments and designs)\",\n \"pubmed_ids\": [\n 11985949,\n 15229327,\n 20368941,\n 21605360,\n 23308135,\n 23844958,\n 23958406,\n 24282316,\n 24406093,\n 24918414,\n 26477996,\n 27924618,\n 28404503,\n 30688830,\n 30921554,\n 34274952,\n 34556954,\n 38207152\n ]\n },\n {\n \"relationship_id\": \"_2dhcsFISl|cj7G775u_Q\",\n \"concepts\": [\n \"walking\",\n \"obesity\"\n ],\n \"n_findings\": 3,\n \"n_studies\": 2,\n \"direction\": \"+\",\n \"summary\": \"3 ratio findings; 2 above 1, 1 below; 2 with CI above 1, 1 with CI below 1; median ratio 1.96 (mixed increments and designs)\",\n \"pubmed_ids\": [\n 33225582,\n 37056656\n ]\n },\n {\n \"note\": \"The graph's physical activity -> obesity findings are split in sign (the ratio findings mix 'activity predicts lower obesity' with 'inactivity predicts higher obesity', so the sign of a ratio depends on the coding of the exposure). There is no concept for cycling or active commuting; variable search finds a few commuting variables (e.g. 'active commuting to/from school', 'being a non-active commuter').\"\n }\n ],\n \"literature\": [\n {\n \"citation\": \"Flint E, Cummins S, Sacker A (2014). Associations between active commuting, body fat, and body mass index: population based, cross sectional study in the United Kingdom. BMJ 349:g4887.\",\n \"doi\": \"10.1136/bmj.g4887\",\n \"finding\": \"Men commuting by active or public modes had BMI about 1 kg/m2 lower than car commuters; women about 0.7 kg/m2.\"\n },\n {\n \"citation\": \"Celis-Morales CA et al. (2017). Association between active commuting and incident cardiovascular disease, cancer, and mortality. BMJ 357:j1456.\",\n \"doi\": \"10.1136/bmj.j1456\",\n \"finding\": \"Cycle commuting associated with lower CVD incidence and mortality (UK Biobank).\"\n }\n ],\n \"prediction\": {\n \"direction\": \"-\",\n \"magnitude\": \"Unadjusted slope: -2 to -6 percentage points of adult obesity per 10 points of walk-or-bike commute share. Poverty-adjusted: -0.5 to -4 points per 10 points.\",\n \"support_if\": \"Poverty-adjusted slope below 0 with a 95% interval below 0.\",\n \"refute_if\": \"Poverty-adjusted slope at or above 0 with an interval that excludes -0.5.\"\n },\n \"plan\": \"Data: EHDP (github.com/nychealth/EHDP-data, commit 547482b9df589006543e13e8cdf6546e00ed9998): indicator 2415 commuting modes, measures 1304 (bicycle %) + 1307 (walked %), UHF42, ACS 2015-19; indicator 2063 adult obesity %, measure 506, UHF42, mean of 2017, 2018 and 2019 (CHS, self-reported BMI); indicator 103 poverty %, measure 221, UHF42, 2015-19. Model: OLS of obesity % on active-commute % (per 10 points), then adding poverty %; HC3 standard errors; sensitivity adding borough fixed effects and excluding Manhattan. Figure: scatter of UHF neighbourhoods coloured by poverty with the fitted adjusted slope and the slope implied by Flint et al. (about -1 kg/m2 BMI for active commuters, translated to obesity prevalence only qualitatively).\",\n \"alternatives\": [\n \"Income, education and race/ethnicity composition (poverty is the named confounder; others are not measured here)\",\n \"Density and Manhattan: walking to work is common where jobs are close, which is also where incomes are high\",\n \"Reverse causation and selection: people with lower BMI choose to walk or cycle, or choose walkable neighbourhoods\",\n \"Commute mode covers workers only; obesity covers all adults\"\n ],\n \"prior\": 0.5,\n \"registrant\": \"mobility agent (Claude Opus 5.5), for J. Vernon\",\n \"registered_at\": \"2026-11-12T15:25:31-0500\",\n \"sha256\": \"ae99f6c6638ffddcd7df7f9576b3f372e9359d98dd86b546a4c396e9a1ba80ae\"\n}",
"statement": "Across New York City's UHF42 neighbourhoods, a higher share of workers who walk or cycle to work goes with a lower share of adults with obesity, after adjusting for neighbourhood poverty."
}History in this collection
- addedv1.0.0Nov 14, 2026this version