Highlights d Mouse S1 and S2 encode overlapping information during a tactile working memory task d Recall responses of previous stimuli are more prevalent in S2 and are relayed to S1 d Category information in S1, but not S2, is necessary for task performance d Network properties of S2 allow task information to persist across behavior states

PURPOSE Nonadherence is a significant issue in cancer care, especially as more oral therapies become available. Measuring and optimizing adherence to such therapies is challenging. In this study, we tested a novel technology that records real-time medication-taking behavior from a smart prescription bottle and can communicate with patients via text message to intervene in cases of nonadherence. METHODS We conducted a 28-patient pilot study to assess the feasibility of this technology in measuring and improving adherence in patients taking capecitabine, an oral chemotherapy agent with a complex, cyclical regimen. The study had a preintervention stage, during which patients were monitored, and an intervention stage, during which the text messaging intervention was enabled. RESULTS During preintervention, patients had an average self-adherence of 89%, and during post intervention, they had an average adherence of 90%. We defined three categories of patients by change in adherence: category 1 (> 8%), category 2 (−8% to 8%), and category 3 (< −8%). Patients in category 1 tended to live in regions with lower average household income (mean = $58,937 in US dollars [USD]) than those in category 2 (mean = $77,482 USD) and category 3 (mean = $90,972 USD). Of poststudy survey respondents, most indicated that they would want to continue using this technology and that they would recommend it to others. CONCLUSION This novel technology is able to monitor, measure, and intervene for patients taking capecitabine in real time. Adherence overall was high, and some patients appeared to benefit more from text-message interventions. Future work should focus on patients deemed high risk for nonadherence.

286 Background: Non-adherence is an important issue in cancer care as more oral cytotoxic and targeted agents become available. Although oral therapies may be more convenient for patients, measuring and optimizing adherence is challenging. The Nomi system records real-time medication taking behavior from a "smart" prescription bottle and displays the data on a web-based interface. Nomi can also communicate with patients via text message to intervene in cases of non-adherence. We report the results of a 28-patient pilot study aiming to assess Nomi’s ability to assist patients taking capecitabine, an oral chemotherapy agent with a complex, cyclical regimen. Methods: Eligible patients were prescribed capecitabine for breast, colorectal, pancreatic, or biliary cancer. The study had a pre-intervention stage, during which patients were monitored, and an intervention stage, in which the text messaging feature was enabled. Adherence was defined as the number of correct doses (both timing and quantity) over the total number of prescribed doses. Conversions were events in which patients took a dose after receiving a text intervention (from Nomi). We defined three categories of patients by percent change in adherence: category 1 ( > 8%), category 2 (-8% to 8%), and category 3 ( < -8%). Results: We collected data from 28 patients (24 pre/post and 4 pre-only). On average, patients were 84% adherent (N = 28; SD = 11%). During pre-intervention, patients had a self-adherence of 89% (SD = 12%), and afterwards, they had an average adherence of 90% (SD = 6%). Most of the patients in category 1 demonstrated a substantial conversion rate ( > 35%). Patients in category 1 tended to live in regions with lower average household income (Mean = $58,937) than those in category 2 (Mean = $77,482) and category 3 (Mean = $90,972). Of survey respondents, 56% indicated that they would want to continue using Nomi, while 67% indicated that they would recommend it to others. Conclusions: This innovative technology is able to monitor, measure and intervene for patients taking capecitabine in real-time. Adherence overall was high, and some patients appeared to benefit more from text message interventions. Future work should focus on patients deemed high risk for non-adherence.

Background: Non-adherence is an important issue in cancer care as more oral cytotoxic and targeted agents become available. Although oral therapies may be more convenient for patients, measuring and optimizing adherence is challenging. The Nomi system records real-time medication taking behavior from a "smart" prescription bottle and displays the data on a web-based interface. Nomi can also communicate with patients via text message to intervene in cases of non-adherence. Aims: The objective is to report the results of a 28-patient pilot study aiming to assess Nomi's ability to assist patients taking capecitabine, an oral chemotherapy agent with a complex, cyclical regimen. Methods: Eligible patients were prescribed capecitabine for breast, colorectal, pancreatic, or biliary cancer. The study had a pre-intervention stage, during which patients were monitored, and an intervention stage, in which the text messaging feature was enabled. Adherence was defined as the number of correct doses (both timing and quantity) over the total number of prescribed doses. Conversions were events in which patients took a dose after receiving a text intervention (from Nomi). Adherence throughout the study was calculated from the data that the bottles collectedwe calculated adherence scores for each patient, during each cycle and study period (pre vs. post-intervention), defined as the number of correct doses (both timing and quantity) over the total number of prescribed doses. We defined three categories of patients by percent change in adherence: category 1 (>8%), category 2 (À8% to 8%), and category 3 (< À8%). Results: We collected data from 28 patients (24 pre/post and 4 pre-only). On average, patients were 84% adherent (N ¼ 28; SD ¼ 11%). During pre-intervention, patients had a self-adherence of 89% (SD ¼ 12%), and afterwards, they had an average adherence of 90% (SD ¼ 6%). Most of the patients in category 1 demonstrated a substantial conversion rate (> 35%). Patients in category 1 tended to live in regions with lower average household income (Mean ¼ $58,937) than those in category 2 (Mean ¼ $77,482) and category 3 (Mean ¼ $90,972). Of survey respondents, 56% indicated that they would want to continue using Nomi, while 67% indicated that they would recommend it to others Conclusions: This innovative technology is able to monitor, measure and intervene for patients taking capecitabine in real-time. Adherence overall was high, and some patients appeared to benefit more from text message interventions. Future work should focus on patients deemed high risk for non-adherence.