Safe Decisions Research Articles

Adaptive Decision-Making for Automated Vehicles Under Roundabout Scenarios Using Optimization Embedded Reinforcement Learning

The roundabout is a typical changeable, interactive scenario in which automated vehicles should make adaptive and safe decisions. In this article, an optimization embedded reinforcement learning (OERL) is proposed to achieve adaptive decision-making under the roundabout. The promotion is the modified actor of the Actor-Critic framework, which embeds the model-based optimization method in reinforcement learning to explore continuous behaviors in action space directly. Therefore, the proposed method can determine the macroscale behavior (change lane or not) and medium-scale behaviors of desired acceleration and action time simultaneously with high sample efficiency. When scenarios change, medium-scale behaviors can be adjusted timely by the embedded direct search method, promoting the adaptability of decision-making. More notably, the modified actor matches human drivers' behaviors, macroscale behavior captures the human mind's jump, and medium-scale behaviors are preferentially adjusted through driving skills. To enable the agent adapts to different types of the roundabout, task representation is designed to restructure the policy network. In experiments, the algorithm efficiency and the learned driving strategy are compared with decision-making containing macroscale behavior and constant medium-scale behaviors of the desired acceleration and action time. To investigate the adaptability, the performance under an untrained type of roundabout and two more dangerous situations are simulated to verify that the proposed method changes the decisions with changeable scenarios accordingly. The results show that the proposed method has high algorithm efficiency and better system performance.

Batteries State of Health Estimation via Efficient Neural Networks With Multiple Channel Charging Profiles

The prognostics and health management (PHM) plays the main role to handle the risk of failure before its occurrence. Next, it has a broad spectrum of applications including utility networks, energy storage systems (ESS), etc. However, an accurate capacity estimation of batteries in ESS is mandatory for their safe operations and decision making policy. ESS comprises of different storage mechanisms such as batteries, capacitors, etc. Consequently, the measurement of different charging profiles (CPs) has a strong relation to battery capacity. These profiles include temperature (T), voltage (V), and current (I) where the CPs patterns vary as the battery ages with cycles. Consequently, estimating a battery capacity, the conventional methods practice single channel charging profile (SCCP) and hop multiple channel CPs (MCCPs) that cause incorrect battery health estimation. To tackle these issues, this article proposes MCCPs based battery management system (BMS) to estimate batteries health/capacity through the deep learning (DL) concept where the patterns in these CPs are changed as the battery ages with time and cycles. Thus, we deeply investigate both machine learning (ML) and DL based methods to provide a concrete comparative analysis of our method. The adaptive boosting (AB) and support vector regression (SVR) are widely compared with long short-term memory (LSTM), multi-layer perceptron (MLP), bi-directional LSTM (BiLSTM), and convolutional neural network (CNN) to attain the appropriate approach for battery capacity and state of health (SOH) estimation. These approaches have a high learning capability of inter-relation between the battery capacity and variation in CPs patterns. To validate and verify the proposed technique, we use NASA battery dataset and experimentally prove that BiLSTM outperforms all the approaches and obtains the smallest error values for MAE, MSE, RMSE, and MAPE using MCCPs compared to SCCP.

Children’s fear in traffic and its association with pedestrian decisions

Introduction: Research on risk for child pedestrian injury risk focuses primarily on cognitive risk factors, but emotional states such as fear may also be relevant to injury risk. The current study examined children's perception of fear in various traffic situations and the relationship between fear perception and pedestrian decisions. Method: 150 children aged 6–12-years old made pedestrian decisions using a table-top road model. Their perceived fear in the pedestrian context was assessed. Results: Children reported greater emotional fear when they faced quicker traffic, shorter distances from approaching traffic, and red rather than green traffic signals. Children who were more fearful made safer pedestrian decisions in more challenging traffic situations. However, when the least risky traffic situation was presented, fear was associated with more errors in children’s pedestrian decisions: fearful children failed to cross the street when they could have done so safely. Perception of fear did not vary by child age, although safe pedestrian decisions were more common among the older children. Conclusions: Children’s emotional fear may predict risk-taking in traffic. When traffic situations are challenging to cross within, fear may appropriately create safer decisions. However, when the traffic situation is less risky, feelings of fear could lead to excessive caution and inefficiency. Practical applications: Child pedestrian safety interventions may benefit by incorporating activities that introduce realistic fear of traffic risks into broader safety lessons.

Health Monitoring System for Autonomous Vehicles using Dynamic Bayesian Networks for Diagnosis and Prognosis

Autonomous Vehicles have the potential to change the urban transport scenario. However, to be able to safely navigate autonomously they need to deal with faults that its components are subject to. Therefore, Health Monitoring System is a essential component of the autonomous system, since allows Fault Detection and Diagnosis. In addition, Prognosis System is also important, since it allows predictive maintenance and safer decisions during vehicle navigation. This paper presents a Hierarchical Component-based Health Monitoring System with Fault Detection, Diagnosis and Prognosis using Dynamic Bayesian Network (DBN) with residue generation, a combination of knowledge-based and model-based detection, diagnosis and prognosis approaches. We evaluate the proposed Dynamic Bayesian Network using different machine learning metrics and a dataset with sensor readings gathered using the CaRINA II autonomous vehicle platform, and the CARLA simulator. Both simulated and experimental results demonstrated a positive performance of the DBNs even with high rate of missing data for some of the model’s variables.

COVID-19: concern about interrupting social isolation of healthcare workers and professionals. What should be done with the results of the available COVID-19 diagnostic tests?

Management of SARS-CoV-2 requires safe decision-making to minimize contamination. Healthcare workers and professionals in confined areas are affected by the risk of the activity and the environment. Isolation of contaminated workers and healthcare professionals requires clinical and diagnostic criteria. On the other hand, interrupting the isolation of healthcare employees and professionals is critical because diagnostic tests do not support clinical decisions. In addition to defining the best test in view of its accuracy, it is necessary to consider aspects such as the stage of the disease or cure, the viral load and the individual's own immunity. Uncertainty about natural and herd immunity to the disease leads to the development of appropriate antivirals, diagnostic tests and vaccines.

The Safety Assessment of MV. Winposh Rampat at Towing Rig Activity during Towing Rig Coslboss at Bintuni Bays

Rig Move is the activity of dismantling and moving drilling rigs by using tugboats from one location to another. There are many conditions and actions during the activity resulting in the ship damage or injury to the crew. In order to know the dangers that will be caused in the rig move activity therefore the safety assessment must be carried out properly. The author identifies the causes of the hazards that occur using the Multiple Causation Model theory. The Multiple Causation Model Theory based on an accident formed due to unsafe conditions and actions. Based on the results of the research, after the causes are identified, the safety assessment can be carried out to look for things that have intolerable properties. With this safety assessment, the captain, as the leader of the ship and the crew in the rig move, knows and are careful in any unsafe conditions and actions that can cause danger and make quick, effective and safe decisions.

An Integrated Framework of Decision Making and Motion Planning for Autonomous Vehicles Considering Social Behaviors

This paper presents a novel integrated approach to deal with the decision making and motion planning for lane-change maneuvers of autonomous vehicle (AV) considering social behaviors of surrounding traffic occupants. Reflected by driving styles and intentions of surrounding vehicles, the social behaviors are taken into consideration during the modelling process. Then, the Stackelberg Game theory is applied to solve the decision-making, which is formulated as a non-cooperative game problem. Besides, potential field is adopted in the motion planning model, which uses different potential functions to describe surrounding vehicles with different behaviors and road constrains. Then, Model Predictive Control (MPC) is utilized to predict the state and trajectory of the autonomous vehicle. Finally, the decision-making and motion planning is then integrated into a constrained multi-objective optimization problem. Three testing scenarios considering different social behaviors of surrounding vehicles are carried out to validate the performance of the proposed approach. Testing results show that the integrated approach is able to address different social interactions with other traffic participants, and make proper and safe decisions and planning for autonomous vehicles, demonstrating its feasibility and effectiveness.

Management of Acute Coronary Syndromes During the Coronavirus Disease 2019 Pandemic: Deviations from Guidelines and Pragmatic Considerations for Patients and Healthcare Workers

Coronavirus disease 2019 (COVID-19) is forcing cardiology departments to rapidly adapt existing clinical guidelines to a new reality and this is especially the case for acute coronary syndrome pathways. In this focused review, the authors discuss how COVID-19 is affecting acute cardiology care and propose pragmatic guideline modifications for the diagnosis and management of acute coronary syndrome patients, particularly around the appropriateness of invasive strategies as well as length of hospital stay. The authors also discuss the use of personal protective equipment for healthcare workers in cardiology. Based on shared global experiences and growing peer-reviewed literature, it is possible to put in place modified acute coronary syndrome treatment pathways to offer safe pragmatic decisions to patients and staff.

An Exploratory Study Investigating COVID-19 Health Safety Measures Using the ARB COVID-19 Restaurant Safety Measure Questionnaire

To obtain important COVID-19 health safety measures information for at risk populations by using an Access Ratings for Buildings based questionnaire. Mixed-methods research design. Questionnaire surveys were conducted over the telephone. Sixteen different restaurants located in Milwaukee, Wisconsin. A verbally administered questionnaire developed by the Rehabilitation Research Design & Disability Center and derived from the Access Ratings for Buildings taxonomy that focuses on restaurant’s COVID-19 health safety measures such as personal protective equipment and CDC recommendations. The ARB COVID-19 Restaurant Safety Measure Questionnaire, which consisted of 11 questions examining the equipment, supplies and policies in place to protect patrons and employees. Out of twenty restaurants, sixteen completed the survey, three were unable to be reached and one refused to participate. Three of the restaurants referred the data collectors to their HR department or restaurant group for questioning. Of the sixteen restaurants that responded, 63% offered dine in services and 88% takeout. Twelve of the restaurants did not offer face masks for patrons, although thirteen did have hand sanitizer for patron use and eleven offered contactless payment methods. 81% of restaurants required employees to wear face masks, 94% had disposable gloves available for employees and 38% did temperature checks on their employees. Fifteen stated they conduct specific sanitation procedures for COVID-19 and believe they are following CDC precautions, although only four currently follow the guidelines. This study provided valuable information that can be used to benefit at risk populations in making informed, safe decisions when deciding where to eat. It also demonstrated the need for restaurants to implement stricter COVID-19 safety measures.

Comparing the usefulness of real-time driving aids in a connected environment during mandatory and discretionary lane-changing manoeuvres

Lane-changing manoeuvre is one of the risky manoeuvres performed by drivers either to reach the planned destination (i.e., mandatory lane-changing; MLC) or to achieve better driving conditions (i.e., discretionary lane-changing; DLC). Essentially both lane-changing types require the driver to acquire surrounding traffic information for efficient and safe lane-changing decisions. However, this does not discount the fact that both these lane-changings are fundamentally different from each other as the urgency of lane-changing is much higher during MLC compared to DLC. Real-time driving aids in a connected environment promise to assist during the lane-changing decision-making process, but the differential effectiveness (or usefulness) of real-time driving aids in a connected environment for these two lane-changing types remains unexplored due to the novelty of a connected environment and the consequent scarcity of data. To fill this research gap, this study collected lane-changing data from 78 participants who performed MLC and DLC in the CARRS-Q Advanced Driving Simulator. Participants were asked to drive in three randomised driving conditions: baseline condition (without driving aids), real-time driving aids in a connected environment with perfect communication, and delayed driving aids in a connected environment. While surrogate measures of safety are analysed and compared using descriptive statistics, a hybrid framework of data mining and classical statistical modelling is employed to examine the usefulness of the real-time driving aids in a connected environment for two lane-changing types. We find that the crash risk associated with MLC is significantly reduced in the driving conditions providing driving aids compared to that of DLC. Results also reveal that the probability of engaging in a hard-braking event decreases for both the lane-changing types in the presence of real-time driving aids in the connected environment driving conditions, but a higher decrease in magnitude is found for MLC. Age and gender-related differential impact have been observed where young and male drivers have a higher possibility of engaging in a hard-braking event when driving without driving aids, but the presence of real-time driving aids reduces such risk. This study concludes that the usefulness (or effectiveness) of the real-time driving aids in a connected environment is a function of the urgency of a task, which is evidently higher during MLC, thus providing the maximum advantage during MLC.

Deep Reinforcement Learning Enabled Decision-Making for Autonomous Driving at Intersections

Road intersection is one of the most complex and accident-prone traffic scenarios, so it’s challenging for autonomous vehicles (AVs) to make safe and efficient decisions at the intersections. Most of the related studies focus on the solution to a single scenario or only guarantee safety without considering driving efficiency. To address these problems, this study proposed a deep reinforcement learning enabled decision-making framework for AVs to drive through intersections automatically, safely and efficiently. The mapping relationship between traffic images and vehicle operations was obtained by an end-to-end decision-making framework established by convolutional neural networks. Traffic images collected at two timesteps were used to calculate the relative velocity between vehicles. Markov decision process was employed to model the interaction between AVs and other vehicles, and the deep Q-network algorithm was utilized to obtain the optimal driving policy regarding safety and efficiency. To verify the effectiveness of the proposed decision-making framework, the top three accident-prone crossing path crash scenarios at intersections were simulated, when different initial vehicle states were adopted for better generalization capability. The results showed that the developed method could make AVs drive safely and efficiently through intersections in all of the tested scenarios.

Detecting, analysing, and modelling failed lane-changing attempts in traditional and connected environments

Lane-changing is a routine yet a complex driving task that has several negative impacts on both traffic flow efficiency and road safety, and thus, lane-changing models have become an indispensable part of microsimulation tools. The existing models only consider lane-changing manoeuvres that are successfully completed while failed lane-changing attempts (i.e., a lane-changing manoeuvre that is aborted after its initiation) are by and large ignored during calibration and validation of lane-changing models. This ignorance leads to structural incompleteness of lane-changing models. In addition, compared with successful lane-changing manoeuvres, failed lane-changing attempts are more likely to disrupt traffic flow and create safety hazards in both the current lane and the target lane, and thus, further warranting its consideration during lane-changing modelling. A connected environment can minimise these adverse effects by providing driving messages about surrounding traffic and subsequent gaps available in the target lane that drivers can utilise to make informed and safe lane-changing decisions. As such, this study investigates the impact of a connected environment on failed lane-changing attempts and addresses the issue of structural incompleteness of lane-changing models using three steps. First, a Wavelet Transform (WT)-based method is employed to detect failed lane-changing attempts from the real data. Second, the impact of failed lane-changing attempts is examined on both traffic flow efficiency and safety parameters such as average speed reduction and Deceleration Rate to Avoid a Crash (DRAC). Moreover, how a connected environment influences these parameters is also explored using a random parameters binary logistic model. Finally, failed lane-changing attempts are incorporated into the existing lane-changing models. At the first step, the WT-based method shows a reasonable accuracy in detecting failed lane-changing attempts when applied to NGSIM dataset and the driving simulator data collected in this study whereby drivers drove the CARRS-Q Advanced Driving Simulator and failed to complete a lane-changing manoeuvre in two randomised driving conditions: baseline (without driving messages) and connected environment (with driving messages). At the second step, we find that failed lane-changing attempts cause a higher speed reduction in both the current lane and the target lane compared to the successful ones. Similarly, a higher DRAC rate is required during failed lane-changing attempts compared to the successful lane-changing attempts, implying a higher crash risk during failed lane-changing attempts. Furthermore, the connected environment has shown to reduce not only the frequency of failed lane-changing attempts but also their negative impacts on surrounding traffic. Moreover, although the developed random parameters model reveals a significant heterogeneity at the individual level, suggesting that while the majority of the drivers tend to abort lane-changing manoeuvres with increase in the relative speed, this does not hold for a small portion of drivers. Finally, by incorporating failed lane-changing attempts into the existing lane-changing models, the predictive accuracy and realism of the lane-changing models have been enhanced.

Cognitive bias: how understanding its impact on antibiotic prescribing decisions can help advance antimicrobial stewardship

The way clinicians think about decision-making is evolving. Human decision-making shifts between two modes of thinking, either fast/intuitive (Type 1) or slow/deliberate (Type 2). In the healthcare setting where thousands of decisions are made daily, Type 1 thinking can reduce cognitive load and help ensure decision making is efficient and timely, but it can come at the expense of accuracy, leading to systematic errors, also called cognitive biases. This review provides an introduction to cognitive bias and provides explanation through patient vignettes of how cognitive biases contribute to suboptimal antibiotic prescribing. We describe common cognitive biases in antibiotic prescribing both from the clinician and the patient perspective, including hyperbolic discounting (the tendency to favour small immediate benefits over larger more distant benefits) and commission bias (the tendency towards action over inaction). Management of cognitive bias includes encouraging more mindful decision making (e.g., time-outs, checklists), improving awareness of one's own biases (i.e., meta-cognition), and designing an environment that facilitates safe and accurate decision making (e.g., decision support tools, nudges). A basic understanding of cognitive biases can help explain why certain stewardship interventions are more effective than others and may inspire more creative strategies to ensure antibiotics are used more safely and more effectively in our patients.

Multi-Agent Safe Policy Learning for Power Management of Networked Microgrids

This article presents a supervised multi-agent safe policy learning (SMAS-PL) method for optimal power management of networked microgrids (MGs) in distribution systems. While unconstrained reinforcement learning (RL) algorithms are black-box decision models that could fail to satisfy grid operational constraints, our proposed method considers AC power flow equations and other operational limits. Accordingly, the training process employs the gradient information of operational constraints to ensure that the optimal control policy functions generate safe and feasible decisions. Furthermore, we have developed a distributed consensus-based optimization approach to train the agents' policy functions while maintaining MGs' privacy and data ownership boundaries. After training, the learned optimal policy functions can be safely used by the MGs to dispatch their local resources, without the need to solve a complex optimization problem from scratch. Numerical experiments have been devised to verify the performance of the proposed method.

Offloading Autonomous Driving Services via Edge Computing

A key challenge for autonomous driving is to process a massive amount of sensor data and make safe and reliable decisions in real time. However, autonomous vehicles often have insufficient onboard resources to provide the required computation capacity. To address this problem, this article advocates a novel approach to offload computation-intensive autonomous driving services to roadside units and cloud for swift executions. Our approach combines an integer linear programming (ILP) formulation for offline optimization of the scheduling strategy and a fast heuristics algorithm for online adaptation. We verify our technique with both synthetic task graphs and real-world deployment. The experimental results show that our approach can improve system performance effectively.

Effect of a brief scenario-tailored educational program on parents' risk knowledge, perceptions, and decisions to administer prescribed opioids: a randomized controlled trial.

This randomized, controlled trial evaluated whether a brief educational program (ie, Scenario-Tailored Opioid Messaging Program [STOMP]) would improve parental opioid risk knowledge, perceptions, and analgesic efficacy; ensure safe opioid use decisions; and impact prescription opioid use after surgery. Parent-child dyads (n = 604) who were prescribed an opioid for short-term use were randomized to routine instruction (Control) or routine plus STOMP administered preoperatively. Baseline and follow-up surveys assessed parents' awareness and perceived seriousness of adverse opioid effects, and their analgesic efficacy. Parents' decisions to give an opioid in hypothetical scenarios and total opioid doses they gave to children at home were assessed at follow-up. Scenario-Tailored Opioid Messaging Program parents gained enhanced perceptions of opioid-related risks over time, whereas Controls did not; however, risk perceptions did not differ between groups except for addiction risk. Scenario-Tailored Opioid Messaging Program parents exhibited marginally greater self-efficacy compared to Controls (mean difference vs controls = 0.58 [95% confidence interval 0.08-1.09], P = 0.023). Scenario-Tailored Opioid Messaging Program parents had a 53% lower odds of giving an opioid in an excessive sedation scenario (odds ratio 0.47 [95% confidence interval 0.28-0.78], P = 0.003), but otherwise made similar scenario-based opioid decisions. Scenario-Tailored Opioid Messaging Program was not associated with total opioid doses administered at home. Instead, parents' analgesic efficacy and pain-relief preferences explained 7%, whereas child and surgical factors explained 22% of the variance in opioid doses. Scenario-tailored education enhanced parents' opioid risk knowledge, perceptions, and scenario-based decision-making. Although this may inform later situation-specific decision-making, our research did not demonstrate an impact on total opioid dosing, which was primarily driven by surgical and child-related factors.

Safety of Autonomous Vehicles

Autonomous vehicle (AV) is regarded as the ultimate solution to future automotive engineering; however, safety still remains the key challenge for the development and commercialization of the AVs. Therefore, a comprehensive understanding of the development status of AVs and reported accidents is becoming urgent. In this article, the levels of automation are reviewed according to the role of the automated system in the autonomous driving process, which will affect the frequency of the disengagements and accidents when driving in autonomous modes. Additionally, the public on-road AV accident reports are statistically analyzed. The results show that over 3.7 million miles have been tested for AVs by various manufacturers from 2014 to 2018. The AVs are frequently taken over by drivers if they deem necessary, and the disengagement frequency varies significantly from 2 × 10−4 to 3 disengagements per mile for different manufacturers. In addition, 128 accidents in 2014–2018 are studied, and about 63% of the total accidents are caused in autonomous mode. A small fraction of the total accidents (∼6%) is directly related to the AVs, while 94% of the accidents are passively initiated by the other parties, including pedestrians, cyclists, motorcycles, and conventional vehicles. These safety risks identified during on-road testing, represented by disengagements and actual accidents, indicate that the passive accidents which are caused by other road users are the majority. The capability of AVs to alert and avoid safety risks caused by the other parties and to make safe decisions to prevent possible fatal accidents would significantly improve the safety of AVs. Practical applications. This literature review summarizes the safety-related issues for AVs by theoretical analysis of the AV systems and statistical investigation of the disengagement and accident reports for on-road testing, and the findings will help inform future research efforts for AV developments.

The role of complementary medicine health literacy in health communication and decision-making

Abstract Background Globally, complementary medicine (CM) product use is high-including vitamins, minerals and herbal medicines. The overuse of CM raises concerns about medicine sustainability that are critical considerations for human and planetary health. Poor CM health literacy may lead to the unnecessary use of CM. This is the first study to explore associations between CM use, CM health literacy, and concurrent use of CM with conventional health care in an Australian population with a view to identifying priority areas for intervention. Methods Cross-sectional online study of Australian adults (N = 2019). A 50-item survey measured demographic and health characteristics, health care and treatment use, medicine use disclosure to health care providers, and CM health literacy. Descriptive statistics, Spearman's correlations and Chi-square analysis were used to analyse preliminary findings. Results In the previous year 49.1% of participants used a CM product. A greater proportion of those who used CM products (88.7%, <0.001), compared to those who did not, used a prescription or over-the-counter pharmaceutical. Lower levels of CM health literacy were associated with herbal supplement use (not vitamin/mineral supplement use) and with disclosing CM use to a medical doctor or pharmacist (p < 0.01). Those who consulted with a GP, specialist or pharmacist, who self-prescribed CM, or used both CM and pharmaceutical medicines had better CM health literacy (p < 0.001) than those who did not. Conversely, those who consulted a CM practitioner had lower CM health literacy than those who did not (p < 0.001). Conclusions The findings suggest that people with lower CM health literacy may be unsure about using CM products, while those with better CM health literacy may have more self-efficacy for making treatment decisions. Interventions are needed to improve public CM health literacy and encourage shared decision-making to ensure safe, effective and sustainable treatment decisions. Key messages People with higher levels of CM health literacy appear to have greater self-efficacy for making treatment decisions, while those with lower CM health literacy are more likely to seek treatment advice. Interventions are needed to ensure people using CM products are communicating with health practitioners about their treatment choices to facilitate safe, effective and sustainable treatment decisions.

Assessing Decision Support Needs of those in Mental Health Crisis Utilizing Cognitive Task Analysis

Extended Abstract Previous research has shown that mental health diagnoses can impact cognitive functioning and decision-making abilities (Bosaipo et al 2017; Stergiopoulos et al 2015; Bora and Pantelis 2015; Rock et al 2014). Decision making difficulties can be particularly problematic when individuals in mental health crises are faced with decisions about their safety and wellbeing. Despite research establishing cognitive and decision-making difficulties for this population, there is an absence in the literature concerning decision support needs of those individuals. Human factors professionals possess a unique set of methods and skills to address these needs and support individuals to make healthier, safer decisions using the latest technologies. This research was part of a larger project aimed at design of a decision support based mobile application that would support individuals during mental health crises. Other phases of the research will include additional user requirements gathering, iterative prototyping, heuristic evaluation, and user testing.

Older Adult Shopping Persona Types and its Impact on Over-the-Counter Medication Misuse

Older adult selection and use of over-the-counter (OTC) medications is informed by a range of motivations and rationales-forming different older adult personas. Holden et al. (2019) categorized older adults seeking OTC medications into two personas: habit followers and deliberators. The goal of this paper is to 1) operationalize and expand on the persona types proposed by Holden et al. 2) use these definitions to characterize the types of older adult personas and 3) explore the relationship between over the counter medication misuse and persona type. Our preliminary analysis identified 3 persona types with differing rates of misuse: 1) deliberators (m=1.8 (range 0-3) instances of misuse), 2) habit followers- formal source dependent (m=0 instances of misuse), 3) habit followers- informal source dependent (m=2.1 (range 1-3) instances of misuse). Information about safe medication choices, provided by physician or pharmacist recommendation, was seen to be critical in safe decision-making about OTC medication selection and use.