Food Shopping Trip Characteristics Before and After the Light Rail

Abstract

Background: The Seattle Food Action Plan recommends improving healthy food access via non-auto transportation options, and light rail is a public transportation system recently introduced and currently being invested in and expanded. When studying food access, time is both an area-based measure of proximity and a component of travel cost to the individual, but proximity does not necessarily correspond with individual travel patterns. Studies of community food access or time use have not analyzed individual reported food shopping trip characteristics for relationships between time and transportation mode, route, and store type. Objective: The overall aim of this study was to understand whether food shopping trips differ with respect to time and mode, and the effects on these travel factors of new light rail implementation as an environmental intervention on travel behaviors. Methods: This study is a secondary analysis of the Travel Assessment and Community (TRAC) Study, a longitudinal study using the introduction of the light rail in Seattle as an environmental intervention. “Cases” were defined as adults living < 1 airline mile from the to-be-opened light rail stations, while the “controls” lived further away from the stations but in neighborhoods with initially similar built environment and census-based demographic characteristics. Food shopping trips from seven-day travel diaries in 2008-2009 (Phase 1) and then in 2010-2011 (Phase 2) after the light rail was running were identified for people with trips in both phases. The final analysis included food shopping trips from 187 cases and 200 controls totaling 1161 trips in Phase 1 and 1086 trips in Phase 2. Trips were characterized by mode, route, store, and time measures and analyzed descriptively only. Person-level outcomes were number of trips per week, classification by primary shopping mode, and average travel time by mode and per primary mode shopper. Results: Non-auto modes were about 30% of all food trips, with 21.1% by walk/bike and 8.4% by transit. Median transit food trip travel time (43 min) was substantially higher than for other modes (20 min for auto and 17 min for walk/bike). Travel times were similar between routes (home-store-home vs other routes) and between store types (supermarket vs other store) within modes for trips overall, but route and store type percentage differed by mode. In Phase 2, only seven trips were associated with light rail, but these were all taken by cases. Mode distribution was similar between phases and case/control. Percentage of trips to supermarkets vs other stores was lower for cases. Transit travel time was lower in Phase 2 vs 1 (39 vs 50 min). Transit trip combinations were the most affected by phase or case. Person-level data indicated that people in Phase 2 vs 1 took fewer walk/bike trips per week on average (0.52 vs 0.67, P=0.0103). There was an increase in transit use by one scheme of classification as a primary transit shopper in Phase 2 (P=0.0125). The difference between cases and controls was smaller in Phase 2 vs 1 for average walk/bike trip travel time by 6 min. Conclusions: Food shopping travel patterns were found to vary by transportation mode. Transit trip characteristics were impacted the most by phase or case/control, with lower transit travel times in Phase 2. Person-level data indicated changes in walk/bike and transit usage by phase and a differential effect on average walk/bike travel time for cases between phases. Understanding trip characteristics and travel behaviors and how they may be impacted by public transportation systems could have implications for transportation planning and inform strategies to promote food access by transit and walk/bike and reduce demand for automobile use.