Implementation of Body Sensor Systems in Healthcare Practice
Permanent URI for this collection
Browse
Recent Submissions
Item Identifying Opioid Withdrawal Using Wearable Biosensors(2021-01-05) Kulman, Ethan; Venkatasubramanian, Krishna; Chapman, Brittany; Carreiro, StephanieWearable biosensors can be used to monitor opioid use, a problem of dire societal consequence given the current opioid epidemic in the US. Such surveillance can prompt interventions that promote behavioral change. Prior work has focused on the use of wearable biosensor data to detect opioid use. In this work, we present a method that uses machine learning to identify opioid withdrawal using data collected with a wearable biosensor. Our method involves developing a set of machine-learning classifiers, and then evaluating those classifiers using unseen test data. An analysis of the best performing model (based on the Random Forest algorithm) produced a receiver operating characteristic (ROC) area under the curve (AUC) of 0.9997 using completely unseen test data. Further, the model is able to detect withdrawal with just one minute of biosensor data. These results show the viability of using machine learning for opioid withdrawal detection. To our knowledge, the proposed method for identifying opioid withdrawal in OUD patients is the first of its kind.Item Diagnosis of Poisoning Using Probabilistic Logic Networks(2021-01-05) Chary, Michael; Boyer, Edward; Burns, MicheleMedical toxicology is the clinical specialty that treats the toxic effects of substances, be it an overdose, a medication error, or a scorpion sting. The volume of toxicological knowledge has, as with other medical specialties, outstripped the ability of the individual clinician to master and stay current with it. The application of machine learning techniques to medical toxicology is challenging because initial treatment decisions are often based on a few pieces of textual data and rely heavily on prior knowledge. Moreover, ML techniques often do not represent knowledge in a way that is transparent for the physician, raising barriers to usability. Rule-based systems and decision tree learning are more transparent approaches, but often generalize poorly and require expert curation to implement and maintain. Here, we construct a probabilistic logic network to represent a portion of the knowledge base of a medical toxicologist. Our approach transparently mimics the knowledge representation and clinical decision-making of practicing clinicians. The software, dubbed \emph{Tak}, performs comparably to humans on straightforward cases and intermediate difficulty cases, but is outperformed by humans on challenging clinical cases. \emph{Tak} outperforms a decision tree classifier at all levels of difficulty. Probabilistic logic provides one form of explainable artificial intelligence that may be more acceptable for use in healthcare, if it can achieve acceptable levels of performance.Item Deployment of a wearable biosensor system in the emergency department: a technical feasibility study(2021-01-05) Miller, Kristen; Hasdianda, Mohammad Adrian; Jambaulikar, Guruprasad; Baugh, Christopher; Divatia, Shreya; Boyer, Edward; Chai, PeterWearable devices to detect changes in health status are increasingly adopted by consumers, yet hospitals remain slow to assimilate these devices into clinical practice. Despite the clear benefits of capturing clinical information in acutely ill patients, such technology remains difficult to implement in emergency medicine. To improve adoption, barriers must first be removed. In our technical feasibility and acceptability trial, we studied the deployment of a wearable wireless biosensor that collects physiological data. We enrolled 44 adult patients receiving care in an emergency department observation unit. After we consented patients for participation, we applied biosensors to their chest and collected basic demographic and clinical information. We then collected biosensor data on an isolated system and measured patient experience via an exit survey. Throughout this process we documented and studied technical challenges. Overall, the technology was feasible to deploy in the emergency department observation unit and was acceptable to participants. Such technologies have tremendous future operational and clinical implications in settings ranging from emergency to home-care.Item Introduction to the Minitrack on Implementation of Body Sensor Systems in Healthcare Practice(2021-01-05) Carreiro, Stephanie; Lai, Jeffrey; Chai, Peter