Start Your Search

Raza Abidi



Author of

  • +

    OS06 - Innovations to Process Non Digital Data (ID 24)

    • Event: e-Health 2019 Virtual Meeting
    • Type: Oral Session
    • Track:
    • Presentations: 1
    • Coordinates: 5/27/2019, 03:45 PM - 04:45 PM, Area 2
    • +

      OS06.03 - Automated Patient Location Identification in Pediatric Emergency Departments (ID 307)

      Raza Abidi, Faculty of Computer Science, Dalhousie University; Halifax/CA

      • Abstract

      Purpose/Objectives:
      Pain, dehydration and anxiety in children are common paediatric Emergency Department (ED) diagnoses but are not well managed. In particular, long ED wait times are well known to increase the anxiety of the patients. We aim to mitigate the anxiety of patients, as well as their families, by providing personalized and location-specific therapeutic and educational interventions while visiting a pediatric ED. Our focus lies on customizing the content with respect to the child’s current location in the ED as well as the current healthcare task, as they move through the stages of examination, investigation and treatment. Using interactive adventure-based scenarios, we aim to proactively collect data about the child’s condition, reduce the child’s anxiety, and lead them to give more meaningful responses about their condition.


      Methodology/Approach:
      In partnership with a Canadian children’s hospital, we extended a mobile, game-based, e-therapeutic and patient education app with personalized and location-aware features. This app leads children and parents through a series of screens asking questions about the individual, their condition and other related information. Based on their responses, the app invokes a variety of therapeutic protocols (e.g., self-administration of Pedialyte for vomiting) and educational videos. The platform is made accessible to children and their parents using an iPad, to mitigate their anxiety, fear and discomfort while waiting in the hospital ED. By leveraging the child’s current location, as well as detected wait and dwell times, the platform is able to dynamically customize the educational content. We designed an intelligent indoor localization method based on (Bluetooth Low Energy) beacon signals, which detect the relative proximity between the iPad and detected beacons as (immediate, near, far, unknown). Based on these proximities, we applied machine learning methods to create an indoor localization model, which can accurately classify the child’s location by correlating multiple beacon signals.


      Finding/Results:
      Our intelligent indoor localization methods have been implemented and validated in a children hospital ED, where 14 beacons where deployed. Data from 29 locations were collected to build indoor localization models (classifiers). Using a hierarchical clustering approach, our approach supports merging multiple locations into cohesive regions to balance localization accuracy with the fine-graininess of indoor localization. Our indoor localization approach was able to recognize the current location of a child with 79% accuracy on average.


      Conclusion/Implications/Recommendations:
      With the proliferation of smart sensors and devices, this innovative project provides numerous opportunities to deliver personalized and timely location-sensitive services to patients.


      140 Character Summary:
      Indoor localization to personalize a mobile e-therapeutic platform for mitigating anxiety, fear and discomfort in children while waiting in the ED waiting room.

  • +

    OS09 - Internet of Things Enhances Care (ID 16)

    • Event: e-Health 2019 Virtual Meeting
    • Type: Oral Session
    • Track:
    • Presentations: 1
    • Coordinates: 5/27/2019, 05:00 PM - 06:00 PM, Area 2
    • +

      OS09.04 - Mobile Health for Personalized Behavior Modification by Personalized Action Planning (ID 306)

      Raza Abidi, Faculty of Computer Science, Dalhousie University; Halifax/CA

      • Abstract

      Purpose/Objectives:
      Chronic illness is affecting a large number of Canadians, with ca. 16% of the population having a multi-morbidity as shown by a recent report from the Canadian Chronic Disease Indicators framework (2017)1. Behavior plays a significant role, with behaviors such as physical inactivity, unhealthy diet and smoking often causing or exacerbating chronic illnesses. To realize long-term positive health behavior change, we present a personalized behavior modification framework called Engage, which computerizes key constructs from the evidence-based Social Cognitive Theory (SCT). 1https://infobase.phac-aspc.gc.ca/ccdi-imcc/


      Methodology/Approach:
      We present a knowledge-based, action-planning and community-driven approach to maximize key SCT indicators such as knowledge and self-efficacy, guided by a knowledge model computerizing SCT constructs. Our approach formulates behavior modification programs as sequences of short-term action plans, which are personalized to the patient and designed to overcome perceived barriers to long-term behavior change (mastery experience). The knowledge indicator constitutes knowledge on the risks and benefits of (un)healthy behaviors, and is influenced by frequent educational messages tailored to a patient’s current situation, personal barriers and behavioral goals. To maximize the self-efficacy indicator, which measures one’s perceived ability to perform a long-term goal, multiple SCT constructs are leveraged, such as mastery experience, social modeling and social persuasion. By performing similarity analysis and data mining on detailed patient profiles and patient feedback, our approach leverages the experiences of similar patients in the community (e.g., regarding health, social status and physical characteristics) to (1) suggest action plans with a strong likelihood of success; (2) offer motivation to patients by seeing similar patients succeed (social modeling); and (3) encourage them to connect, exchange advice and provide encouragement on barriers to be overcome (social persuasio


      Finding/Results:
      The Engage framework implements a holistic approach to behavior modification, including (a) collecting an up-to-date patient profile and assessing patient-specific SCT indicators; (b) selecting and tailoring a behavior modification program, based on individual patient profiles and collective patient experiences; (c) delivering timely educational and motivational messages; and (d) monitoring patient compliance and aggregating community-wide feedback on behavior modification programs. The framework includes a number of key components: 1) A core back-end service, which keeps the knowledge model, patient profiles and algorithms for similarity analysis and data mining. 2) A front-end web portal, allowing patients to enter their personal profile, fill out questionnaires for weekly monitoring, and selecting between different action plans to perform. 3) A tailored social network that facilitates patients to connect with other similar patients, post their progress, receive encouragement and see others, similar to their individual situation, succeed. 4) A mobile app (Android, iOS) used by patients to submit progress towards their weekly action plan, receive tailored educational and motivational messages, and serving as a portal into the social network.


      Conclusion/Implications/Recommendations:
      The Engage framework currently includes behavior change content for the long-term goal of increasing physical activity – although other content can be easily plugged in – and is undergoing usability tests. We are planning an extensive clinical evaluation of the system, including patients with high risk factors for chronic illness and those suffering from chronic illness.


      140 Character Summary:
      To realize long-term health behavior change, we present a knowledge-based, action-planning and community-driven system guided by the Social Cognitive Theory.

  • +

    OS21 - Big Data Provision for Providers (ID 34)

    • Event: e-Health 2019 Virtual Meeting
    • Type: Oral Session
    • Track:
    • Presentations: 1
    • Coordinates: 5/28/2019, 01:15 PM - 02:15 PM, Area 3
    • +

      OS21.04 - Pathology Data Analytics to Optimize Laboratory Utilization by Interactive Scorecards (ID 146)

      Raza Abidi, Faculty of Computer Science, Dalhousie University; Halifax/CA

      • Abstract

      Purpose/Objectives:
      Pathology laboratories provide service to both primary-care and tertiary care providers, helping with disease diagnosis and therapeutic choices. Typically, physicians request a pathology test order which may contain multiple tests; the operational question is whether these tests are relevant and useful with respect to the patient’s profile. As service demands on pathology laboratories is increasing, there is a realization to streamline the operations with respect to clinical guidelines and local clinical workflows to optimize operational costs whilst improving order relevancy and result accuracy. A Canadian study has shown that physicians test ordering behaviour can be modified by education, personalized audit and peer comparisons. Pathology laboratories generate large volumes of clinical data that can be analyzed to monitor, manage and optimize laboratory utilization. This project aims to optimize pathology laboratory utilization by detecting superfluous (clinically irrelevant, unnecessary, repetitive) lab orders by physicians and then educating physicians about test ordering guidelines. Our specific objectives are: (1) To develop and deploy a Pathology Laboratory Utilization Scorecards (PLUS) platform that offers end-to-end pathology ‘big’ data analytics services to optimize laboratory utilization; (2) To provide primary care physicians personalized laboratory utilization scorecards so that they can examine their test ordering pattern and adjust their test orders accordingly; (3) To provide pathology laboratory managers a live dashboard showing the volume and type of orders to assist them with resource planning; and (4) To generate meaningful order-sets to improve test ordering patterns and guideline compliance.


      Methodology/Approach:
      Big data analytics approach is taken to develop PLUS that hosts a suite of health data analytics tools to (i) standardize pathology data using SNOMED-CT; (ii) integrate pathology data from feeder health information systems (such as ADT, EDS); (iii) analyze lab data using machine learning methods—clustering methods are applied to develop physician order profiles to stratify physicians with respect to their patient case-mix (as opposed to their order type and volumes) for inter-physician peer comparisons, and rule association methods are applied to generate order-sets and to evaluate test orders based on previous order patterns; and (iv) visualize analytical results as interactive scorecards—advance data visualization techniques are used to visualize the multi-dimensional physician scorecard, giving physicians the ability to dynamically interact with their scorecard to get personalized views of their ordering behaviour and comparisons with their peer-group. We performed data analytics on physician’s test orders for the period 2011-2017 with a dataset comprising around 8 million test orders from 200 physicians.


      Finding/Results:
      We have developed PLUS to optimize the pathology laboratory utilization in the central zone (Halifax) that annually performs on average 15 million laboratory tests for 200,000 patients. PLUS is securely web-accessible to physicians to privately audit their ordering profile in terms of volume of test orders with abnormal rates, repetition rates, yearly comparison and comparison to their peers.


      Conclusion/Implications/Recommendations:
      ‘Choosing Wisely Canada’ is promoting sustainable healthcare by optimizing the utilization of healthcare services. This project engages primary care physicians to help optimize lab utilization, and this will impact NSHA annual budget by reducing demand for diagnostic services whilst increasing patient safety in line with Choosing Wisely principles.


      140 Character Summary:
      A big health data analytics platform using artificial intelligence to analyze pathology lab data to optimize pathology lab utilization and increase patient safety

  • +

    OS29 - Virtual Care in Mental Health (ID 50)

    • Event: e-Health 2019 Virtual Meeting
    • Type: Oral Session
    • Track:
    • Presentations: 1
    • Coordinates: 5/29/2019, 10:30 AM - 12:00 PM, Room 2
    • +

      OS29.03 - A Mobile Early Stimulation Program Supporting Children with Developmental Delays (ID 184)

      Raza Abidi, Faculty of Computer Science, Dalhousie University; Halifax/CA

      • Abstract

      Purpose/Objectives:
      In Brazil, child developmental delays has became a major public health concern after the Zika virus outbreak. Early Stimulation Program (ESP) is a standardized intervention to treat developmental delays in children (aged 0-3 years) through a series of specialized exercises that help the child achieve the age-specific developmental milestones. Limited ESP are available in specialized therapeutic centers due to lack of healthcare resources. As such, Brazilians face significant challenges to access ESP and as a result several children do not get the required ESP and end up with permanent cognitive impairment. Our objective is to leverage digital health technologies to provide accessible, affordable and personalized ESP that can be administered by the child’s family at home. This objective is pursued by a mobile health application—i.e. BraziLian Early Stimulation System (BLESS)—that offers (a) clinical decision support to assist healthcare professionals to prescribe a personalized ESP to a child with developmental delays; (b) self-management support to enable the child’s family to administer the prescribed ESP in a home-based setting.


      Methodology/Approach:
      To develop personalized ESP, our approach is to implement the developmental milestones advocated by WHO and the International Classification of Functioning and Disabilities. These developmental milestones are being pursued by taking the “prepared-informed-motivated” approach (based on the Innovative Care for Chronic Conditions) to educate parents to deliver ESP to their child at home. Based on the Brazilian Early Stimulation Guidelines we have developed an Early Stimulation Activities (ESA) database that comprises cognitive development exercises that are classified by age, impairment level and complexity of execution. We have developed a decision logic that guides the healthcare provider to select and personalize ESA to generate a ESP in line with the child’s developmental challenges. To educate the child’s parent on how to perform the ESA, educational videos and messages about the ESA are sent to the parent’s mobile phone as per their child’s prescribed ESP. Knowledge translation strategies to engage the stakeholders (health team, families, and local experts) are pursued.


      Finding/Results:
      BLESS comprises a web-based clinical decision support platform for health professionals and a mobile health app for families. The therapist platform provides assessment, monitoring and management support, helping them to perform standardized child assessment and therapy planning in a shared care planning environment. BLESS offers a comprehensive platform for ESP, including registering patient and family; assessing child’s developmental needs and the family’s efficacy to perform ESP; selecting and personalizing ESP; and monitoring the child’s overall ESP progress. The BLESS mobile app offers ESP educational material in terms of short videos and step-by-step instructions written in the plain language); a diary for capturing child’s progress and monitoring parent’s engagement; and the overview of child’s progress.


      Conclusion/Implications/Recommendations:
      BLESS will be deployed at the Mother and Child healthcare centre in the Northeast of Brazil for a pilot study to assess the intervention impact on child’s development and parent’s engagement. BLESS is an innovative digital health based solution to administer ESP at home to overcome a child’s developmental challenges and help the child lead a normal life.


      140 Character Summary:
      BLESS is an innovative mobile system that provides personalized early stimulation program, to empower and educate parents of children with developmental delays.