with different prompting strategies such as computer generated verbal instructions, light cues highlighting kitchen cabinets, projected cues showing enlarged pictures of recipe ingredients, and smart glass cues allowing users to see through cabinet doors. Main Outcome Measure(s): Statement of unmet needs in kitchen activities, Task completion time as a measure of efficiency of prompting techniques, user preference scores towards different prompting techniques Results: Design of the Smart Cueing Kitchen (SCK) was motivated by user needs such as: Users need assistance in remembering locations of recipe ingredients and keeping track their progress while preparing a recipe; users need to be guided before they realize they have failed and feel frustrated; guidance should be minimally distracting; users should be able to perceive their kitchen as a stress reducer rather than a stress producer. The context aware prompter software uses the information from a portable network of sensors in the kitchen for automatic recognition of user’s activities and adaptively guides users to complete the task at hand using various prompting strategies. The safety monitoring software ensures user’s safety during and after the use of kitchen appliances. For example, appliances such as the stove, oven, and water faucets are continuously monitored and can be automatically switched off by the kitchen software or remotely by the user or family member using a cell phone application. The projected picture prompts were the most preferred. Participants were also most efficient at item retrieval when picture prompts were used and least efficient when guided by computer generated verbal instructions alone. Participants preferred multi modal cues and had strong personal inclinations towards colors of the visual cues and types of computer voices. A second round of ethnographic interviews that is currently being conducted will add further insights and be used to validate earlier findings. Another clinical protocol is currently being developed to evaluate the usability, reliability, and effectiveness of the automated guidance system in contrast to similar commercially available alternatives such as recipe apps on cell phones. This protocol is also aimed at evaluating the clinical utility of the SCK in improving task efficiency and independence. Conclusions: The SCK was built as a cognitive orthosis for people with cognitive impairments to provide support for kitchen activities and take remedial actions to ensure user safety. The SCK software provides multiple customization options to fit the needs and personal preferences of individuals with cognitive impairments. In addition to being an automated cooking assistant and safety monitor the SCK system has promising future applications as a training tool during rehabilitation process. By keeping people with cognitive impairments active and safe in their home envi- ronments and hence away from nursing homes and hospitals, smart kitchen technologies may contribute towards reducing healthcare cost in the long term. Key Words: smart homes, cognitive orthosis, traumatic brain injury, Alz- heimer’s disease Disclosure: Harshal Mahajan has nothing to disclose. Article 4 (NIDRR) Disparity in Access to Healthcare among Individuals with Physical Disabilities: 2001-2010 Elham Mahmoudi (University of Michigan), Michelle A. Meade Objective: This study examines disability-related disparities in access to healthcare and investigates the factors associated with reporting no access to care. Design: A retrospective analysis of reported access to healthcare for adults, 18 or older, with or without physical disabilities, who self- identified as non-Hispanic White (NHW), non-Hispanic African- American (NHB), or Hispanic was conducted. Access to care was measured as 1) whether the individual reported being unable to get medical care, and 2) whether the individual reported being unable to receive dental care, and 3) whether the individual reported being unable to get prescription drugs. We modeled the three measures of access using logistic regressions. Setting: Secondary analysis of cross-sectional data from the Medical Expenditure Panel Survey (2001-2010). Participants: A nationwide sample of adults, 18 or older, with or without physical disabilities, who self-identified as non-Hispanic White (NHW), non-Hispanic African-American (NHB), or Hispanic (secondary data analysis). Interventions: N/A Main Outcome Measure(s): Access to care was measured as 1) whether the individual reported being unable to get medical care, and 2) whether the individual reported being unable to receive dental care, and 3) whether the individual reported being unable to get prescription drugs. Results: We analyzed a total of 138,670 adults (with mild to severe physical disabilitiesZ27,276; without any physical disabilityZ111,404). Our analysis indicated the odds of reporting not getting medical care, dental care, and prescription drugs are 39% (p < 0.001), 51% (p < 0.001), and 38% (p < 0.001) higher for individuals with physical disabilities, respectively. Furthermore, in comparison with Whites with disabilities, our data showed that Hispanics with physical disabilities have an additional 37% (p < 0.015) higher odd of reporting inability to get prescription drugs. Our models of access indicated that being poor (p <0.001), lacking health insurance coverage (p < 0.001), being a smoker (P < 0.001), residing in South (P < 0.007), and being female (p < 0.001) also significantly increase the odds of reporting no access to care. Conclusions: There are large and significant disparities in access to healthcare between adults with and without physical disabilities. Key Words: Physical Disability; Healthcare Disparity; Access to health- care; Healthcare Policy Disclosure: Elham Mahmoudi has nothing to disclose. Article 5 (NIDRR) Time course of kinematic improvements in survivors of stroke during upper-extremity robotic rehabilitation Crystal L. Massie (University of Maryland School of Medicine), Susan Conroy, George Frederick Wittenberg, Jill Whitall, Christopher T. Bever Objective: Determine when improvements in unassisted reaching occur during robotic rehabilitation for chronic survivors of stroke. Design: Randomized control trial Setting: VA Clinical Research Lab Participants: Twenty-four participants with chronic stroke (average 2.7 years post-stroke and 56.8 years old). Participants met inclusion criteria if at least 6 months post-stroke and Fugl-Meyer (FM) score between 7 and 38. Interventions: Participants randomized into robot only (60 mins) or transition to task groups (45 mins robot and 15 minutes of functional task practice). Participants completed 4 weeks of wrist robot training prior to 4 weeks on the planar robot (3 visits per week). Robot provided assistance as needed during therapy but no assistance during assessments. Main Outcome Measure(s): Outcome measures were assessed on the planar robot with unassisted reaches to 8 equally spaced targets around a home target prior-to and during each therapy session. Outcomes included percent of targets hit, movement time (sec), peak velocity (m/s), path length (m), and movement units. Results: No group differences were observed. Improvement in percentage of targets hit was related to baseline FM scores (FM scores 15-20 had greatest gains). The number of targets hit significantly increased and was retained after the 3 rd visit on; movement time and peak velocity signifi- cantly improved and were retained after the 6 th visit. Conclusions: Results inform intervention planning in that additional daily time on robot did not improve unassisted reaching kinematics and most gains were achieved within two weeks. Further, response to planar robot therapy may depend on initial FM scores. e2 Neuroscience Focus www.archives-pmr.org