The Open Collections website will be unavailable July 27 from 2100-2200 PST ahead of planned usability and performance enhancements on July 28. More information here.

Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Examining patient’s mobile phone access and planning a virtual care intervention using mHealth and conversation… Ghaseminejad-Tafreshi, Niloufar 2021

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2021_november_ghaseminejad-tafreshi_niloufar.pdf [ 2.29MB ]
Metadata
JSON: 24-1.0400040.json
JSON-LD: 24-1.0400040-ld.json
RDF/XML (Pretty): 24-1.0400040-rdf.xml
RDF/JSON: 24-1.0400040-rdf.json
Turtle: 24-1.0400040-turtle.txt
N-Triples: 24-1.0400040-rdf-ntriples.txt
Original Record: 24-1.0400040-source.json
Full Text
24-1.0400040-fulltext.txt
Citation
24-1.0400040.ris

Full Text

EXAMINING PATIENT’S MOBILE PHONE ACCESS AND PLANNING A VIRTUAL CARE INTERVENTION USING MHEALTH AND CONVERSATION ANALYTICS by  Niloufar Ghaseminejad-Tafreshi  B.Sc., The University of British Columbia, 2013  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Experimental Medicine)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2021  © Niloufar Ghaseminejad-Tafreshi, 2021  ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis entitled:  Examining Patient’s Mobile Phone Access and Planning a Virtual Care Intervention Using mHealth and Conversation Analytics  submitted by Niloufar Ghaseminejad-Tafreshi in partial fulfillment of the requirements for the degree of Master of Science in Experimental Medicine  Examining Committee: Dr. Richard Lester, Department of Medicine, University of British Columbia Supervisor  Dr. John A Staples, Department of Medicine, University of British Columbia Supervisory Committee Member  Dr. Karen Tran, Department of Medicine, University of British Columbia Additional Examiner  iii  Abstract Introduction: Unplanned hospital readmissions create stress for patients and their families while placing individuals at risk for negative outcomes and increasing healthcare system costs. Development of effective interventions to reduce readmissions involves timely discharge planning, transitional care, and stakeholder uptake. Mobile health (mHealth) and machine learning technology may help improve coordination of care, identify the underlying reasons for complications, and potentially reduce readmissions.  Methods: To determine whether mHealth can help streamline and improve transitional care after discharge from the hospital, we will utilize a two-way text messaging virtual care platform to be piloted at the medical wards in the Vancouver General Hospital (VGH) Clinical Teaching Unit (CTU). Prior to launching the program, we conducted a survey of patients admitted to the CTU to determine mobile phone access, usage, and preferences to better understand the population we wish to serve. Using this information, we designed an mHealth intervention protocol that is patient-centered and collaborative.  Results: We found that a two-way text messaging mHealth platform would likely be well-placed to facilitate better transitional care and to understand the underlying reasons for readmissions. Our survey results indicated that 86% of participants had access to a mobile phone, 63% of whom owned their own device and 23% of whom had access via a proxy (e.g., family or caregiver). These findings indicate that most patients can participate in mHealth interventions that rely on mobile phones and that engaging a proxy may further expand inclusivity. Lastly, we conducted training sessions and consulted with hospital staff to ensure the study protocol meets iv  end-user needs and preferences. Using these findings, we developed a framework that utilizes natural language processing and machine learning to analyze patient text message conversations with their health care provider (HCP).  Conclusion: Our findings suggest that mHealth virtual care platforms are feasible and accessible in a hospital setting, which may help in reducing the burden of hospital readmission on patients, their families, and the health care system.  v  Lay Summary Returning to the hospital and being readmitted soon after discharge can have negative consequences on the patient as well as those who care for them. With the enhancement of technology, mobile phones are now increasingly used in the healthcare field. One way that patients can use technology to access care is through the mobile phone. We developed a program where patients can use their mobile phones to stay connected with their healthcare team after being discharged from the hospital. To learn more about the population we want to help, we conducted a survey to find out if patients would be able and interested in communicating with the healthcare team using their mobile phone. We learned that text-messaging would be the most accessible way for patients to communicate. Finally, we developed a framework to assist the implementation of the study.   vi  Preface This thesis was written by me under the supervision of Dr. Richard Lester. My committee member, Dr. John A Staples, provided guidance, feedback, and revisions.  Some of the contents of this thesis include excerpts from a manuscript on which I am a co-author. However, the inferential statistical analysis and findings are unique to this thesis and performed by me.  Chapter 2 is based on a Patient Phone Access Survey study conducted in the Vancouver General Hospital (VGH) Clinical Teaching Units (CTUs) and led by Maryam AboMoslim, with support from Abdulaa Babili and Samia El Joueidi. This study was approved by the University of British Columbia (UBC) Behavioural Research Ethics Board (certificate H19-03366).   Chapter 3 is based on an interventional pilot study approved by the UBC Clinical Research Ethics Board (certificate H19-02740). I was responsible for ethical approval, study design, stakeholder liaising, training sessions, and post-study survey development. This chapter is also based on a novel medical conversation analytics tool called ConVIScope. Some content from this chapter is drawn from a poster I presented, lead, and co-authored for the Centre for AI Decision-making and Action (CAIDA) conference on November 25th, 2020 and the Experiment Medicine student Expo on May 18th, 2021.   vii   Table of Contents Abstract ............................................................................................................................... iii Lay Summary ......................................................................................................................... v Preface ................................................................................................................................. vi Table of Contents................................................................................................................. vii List of Tables ......................................................................................................................... x List of Figures ....................................................................................................................... xi List of Abbreviations ............................................................................................................ xii Acknowledgements ............................................................................................................ xiii Dedication .......................................................................................................................... xiv Chapter 1: Introduction ..................................................................................................................1 1.1 Problem and Purpose Statement ........................................................................................................... 1 1.2 Unplanned Hospital Readmission .......................................................................................................... 1 1.2.1 History ............................................................................................................................................... 3 1.2.2 Risk Factors and Prediction Models .................................................................................................. 4 1.2.3 Interventions ..................................................................................................................................... 8 1.3 Virtual Care and Mobile Health.............................................................................................................. 9 1.3.1 WelTel .............................................................................................................................................. 11 1.3.2 Natural Language Processing and Machine Learning ..................................................................... 12 1.4 WelTel Within the Context of The Patient’s Medical Home ................................................................ 13 viii  1.5 Research Questions and Objectives ..................................................................................................... 16 Chapter 2: H2home Patient Access Survey .................................................................................... 17 2.1 Background .......................................................................................................................................... 17 2.2 Methods ............................................................................................................................................... 18 2.3 Research Question ............................................................................................................................... 20 2.4 Hypotheses ........................................................................................................................................... 20 2.5 Results .................................................................................................................................................. 21 2.5.1 Study Population ............................................................................................................................. 21 2.5.2 Survey Responses ............................................................................................................................ 25 2.5.3 Accessibility and Usage ................................................................................................................... 27 2.5.4 Age and Sex ..................................................................................................................................... 28 2.5.5 Risk of Readmission Assessment Score (RRAS) ............................................................................... 30 2.6 Discussion ............................................................................................................................................. 32 Chapter 3: H2home mHealth Intervention Pilot Development ....................................................... 35 3.1 Introduction ......................................................................................................................................... 35 3.2 Thesis Objectives .................................................................................................................................. 38 3.3 Research Aims ...................................................................................................................................... 38 3.4 Study Interruption ................................................................................................................................ 39 3.5 Methods ............................................................................................................................................... 39 3.5.1 Patient Consent ............................................................................................................................... 42 3.5.2 Patient Registration ......................................................................................................................... 43 3.6 Analysis Plan ......................................................................................................................................... 46 3.7 Limitations ............................................................................................................................................ 47 3.8 H2home WelTel Framework ................................................................................................................ 47 3.9 ConVIScope Training ............................................................................................................................ 49 ix  3.10 Next steps ............................................................................................................................................. 50 Chapter 4: Conclusion .................................................................................................................. 51 References ........................................................................................................................... 54 Appendices .......................................................................................................................... 68 APPENDIX A: Readmission Risk Assessment Score (RRAS) ................................................................................. 68 APPENDIX B: My Discharge Plan ......................................................................................................................... 69 APPENDIX C: Contingency Table of Patients’ Internet Access by SMS Use  ....................................................... 71 APPENDIX D: Patient Mobile Phone Access Survey ........................................................................................... 74 APPENDIX E: H2home Post-Survey  .................................................................................................................... 35 x  List of Tables Table 2.1 Variables Considered and Method of Assessment to Conduct Statistical Tests........................................... 20 Table 2.2 Survey Participants’ Characteristics ............................................................................................................. 23 Table 2.3 Participant Mobile Phone Access ................................................................................................................. 25 Table 2.4 Participant Mobile Phone Use ...................................................................................................................... 26 Table 2.5 Participant Ranking of Communication Methods by Frequency of Usage ................................................... 26 Table 2.6 Correlation matrix between variables using Fisher’s exact test ................................................................... 28 Table 2.7 One-way Texting with HCP and Risk of Readmission Assessment Score (RRAS) .......................................... 30 Table 2.8 Two-way Texting with HCP and Risk of Readmission Assessment Score (RRAS) ......................................... 30 Table 2.9 Measures of Association for Two-way and One-way Communication with HCP ......................................... 31 Table 3.1 WelTel H2home Features ............................................................................................................................. 40  xi  List of Figures Figure 2.1 Participant Flowchart .................................................................................................................................. 22 Figure 2.2 Most Responsible Diagnosis by Category ................................................................................................... 23 Figure 2.3 Participants’ Neighbourhood of Residence Using First Three Letters of Postal Code ................................. 24 Figure 2.4 Age of Mobile Phone Owners ...................................................................................................................... 29 Figure 2.5 Age of Patients with Internet Access........................................................................................................... 29 Figure 2.6 Age of Patients Using SMS .......................................................................................................................... 29 Figure 3.1 Hospital Discharge and Readmission Cycle ................................................................................................. 37 Figure 3.2 H2home Intervention Pilot .......................................................................................................................... 41 Figure 3.3 Participant Registration (part 1) ................................................................................................................. 44 Figure 3.4 Participant Registration (part 2) ................................................................................................................. 45 Figure 3.5 Participant Registration (part 3) ................................................................................................................. 45 Figure 3.6 H2home WelTel Hospital Readmission Prevention Framework .................................................................. 48 Figure 3.7 Supervised Machine Learning of Patient Test Training Data ...................................................................... 50  xii  List of Abbreviations ADI – Area Deprivation Index BC – British Columbia CFPC – College of Family Physicians of Canada  CIHI – Canadian Institute for Health Information CIMD – Canadian Index of Multiple Deprivation CML – Care Management Lead HCP – Health Care Provider HRRP – Hospital Readmission Reduction Program  IHI – Institute of Healthcare Improvements mHealth – Mobile Health NA – Not Applicable NLP – Natural Language Processing PMH – Patient’s Medical Home SES – socioeconomic status SMS – Short Message Service U.S. – United States VCH – Vancouver Coastal Health VGH – Vancouver General Hospital WHO – World Health Organization     xiii  Acknowledgements I would like to express my deepest gratitude to my supervisor, Dr. Richard Lester, for his mentorship, support, and compassion throughout the past two years. Dr. Lester, thank you for your flexibility, positive attitude, and innovative thinking. Despite the many obstacles presented to us by the global pandemic, you continued to help me move forward.  I would also like to extend my appreciation to Dr. John Staples for his time and guidance. Dr. Staples, thank you for your valuable feedback, expertise, and critical suggestions. Your contribution to my research really helped elevate my thesis.   I would like to thank the entire mHealth Research Group and WelTel team for your support, friendship, advice, and the unforgettable memories.   I would like to express my love and appreciation to my mother, father, brother, family, and friends. Thank you for your unconditional love and support. I feel incredibly fortunate to be surrounded by so many incredible people who inspire me every single day. xiv  Dedication  To my grandmothers.1  Chapter 1: Introduction 1.1 Problem and Purpose Statement One in eleven patients in Canada are readmitted to the hospital within only 30 days of being discharged.1,2 Unplanned hospital readmissions are an indicator of negative health outcomes, associated with patient dissatisfaction, and costly at $2.8 million each year.3–5 In an effort to better understand and reduce unplanned readmission rates, health service researchers have placed significant focus on examining contributing risk factors such as sex, chronic conditions, socio-economic status (SES), and age.1,6,7 However, there is no clear and measurable consensus regarding the reliability of methods used to evaluate or predict a readmission incident.8–10 Shifting the focus towards understanding the underlying reasons behind readmissions might be a better strategy to overcome the existing gaps in care during the vulnerable transition from hospital to community.11 The purpose of this research is to examine approaches designed to reduce the risk of unplanned hospital readmission and to explore the use of a novel mobile health (mHealth) tool with the potential to improve post-discharge follow-up and transitional care. This thesis outlines the planning stages and preliminary assessments of this innovative mHealth pilot intervention.  1.2 Unplanned Hospital Readmission An unplanned hospital readmission refers to when a patient is readmitted to the hospital shortly after discharge, in an unscheduled way. This indicator is calculated as a rate and, depending on the institute, can vary in methodology. The Canadian Institute for Health Information (CIHI) measures it as the percentage of patients returning urgently to the hospital within 30 days following discharge, with adjustments to account for patient heterogeneity (e.g., age, mental 2  health status, and end-of-life care).12 When examining hospital readmissions, it is also important to consider hospital type (e.g., large teaching hospitals vs. small) or hospital location (e.g., urban vs. rural) as more complicated patient cases can be at a higher risk of returning to hospital.13 For instance, for small hospitals in rural areas, there may be fewer services available in the community to ensure quality follow-up care.   Estimates suggest that approximately 27% of readmissions to the hospital are preventable during the index hospitalization10,14 but there are many other factors that have been shown to prompt a patient’s return to the hospital which are difficult to anticipate using current “one-size-fits-all” approaches. Despite increasing attention to hospital readmissions rates, relatively few programs deploy patient-centered interventions as most studies use a cohort-based model and rely heavily on secondary data sources (e.g., administrative data). One issue is the complexity of events triggering readmission and the heterogenous nature of patient experiences; readmissions are often not attributable to a single cause.15 The readmission diagnosis is typically discordant from index admission diagnosis; for example, patients initially admitted for heart failure may be readmitted for acute kidney injury, hyperkalemia, orthostatic syncope, or cellulitis. This makes it difficult to target interventions based on index diagnosis.  Many of the identified risk factors for readmission are not proximate causes of readmission and cannot be modified (e.g., age, index diagnosis, and presence of comorbidities). Furthermore, community characteristics, such as lack of access to transportation, may contribute to variations in hospital readmissions rates in ways that are challenging to address.16 In one study of patients readmitted to the hospital, lower SES patients were less likely to report understanding self-care, 3  have a follow-up appointment scheduled, and be able to fill their prescriptions.17 The patient group at highest risk of readmission may thus differ from the group most likely to benefit from intervention.   One way that policy makers have addressed hospital readmission is by improving discharge planning and coordination of care.5 Discharge planning protocols can differ depending on the country, province, or even health region with implications for both the patients and the hospital.18 Thus, a lack of continuity of care also introduces complex problems for healthcare providers who do not always have access to complete patient information.19 Examining post-discharge continuity of care and choosing the right interventions for the healthcare setting can improve the quality of care and provide individualized care to patients.   Since patients, families, and the care team must communicate effectively to ensure a smooth transition from the hospital back to the community, we need to use modern tools to examine their communication and the patient’s perspective. We propose a new focus on the specific, individual, and proximate causes of readmissions using alternative types of data that provide insight into patient-focused needs. A patient-centered approach can allow the care team to deploy individually tailored, timely interventions that might more efficiently control healthcare costs, improve the quality of patient care, and reduce unplanned hospital readmissions.  1.2.1 History As early as the 1970s, researchers have been interesting in understand the proportion of patients discharged from the hospital in unstable conditions and with poor outcomes. Subsequently, 4  healthcare professionals, particularly nurse specialists, began to pursue collaborative transitional care services and cost-effective interventions.20–23 In 2008, a group of researchers from the United States (U.S.) Institute of Healthcare Improvements (IHI), introduced The Triple Aim required to improve the health care system: (1) improving the experience of care, (2) improving the health of populations, and (3) reducing per capital costs of health care.24 Following introduction of The Triple Aim in 2010, the U.S. Patient Protection and Affordable Care Act enacted the Hospital Readmission Reduction Program (HRRP) which was then implemented in 2013. The goal of the HRRP is to incentivize the improvement of inpatient care quality by imposing penalties on hospitals with “higher-than-expected” rates of readmission for acute myocardial infarction, chronic obstructive pulmonary disease, heart failure, pneumonia coronary artery bypass graft surgery, elective primary total hip arthroplasty, and total knee arthroplasty.25,26 The three Triple Aim dimensions have been used to assess the positive impacts of HRRP on hospital readmission rates. Although there is a debate on the HRRP’s impact and effectiveness, there is evidence that the program benefits both Medicare patients and those across other insurance types.27 However, pay-for-performance and punitive models are not appropriate in the Canadian context given its publicly funded healthcare system.28 In Canada, healthcare leader are taking an approach that focuses on preventive, community-integrated, and interprofessional methods to reducing hospital readmissions, including an increased use of electronic health records, virtual care interventions, and physician financial incentives.29–31     1.2.2 Risk Factors and Prediction Models There are several health conditions that place one at an elevated risk for being readmitted to the hospital soon after discharge. Prolonged hospital stays, comorbidities, and polypharmacy are 5  commonly known patient-level risk factors for 30-day hospital readmission. Furthermore, cardiovascular, renal, and kidney diseases are more common among readmitted patients.32,33 One study found that patients dependent on home care, of poor health, and with regular use of ten or more medications were at increased odds of readmission within 30 days of discharge; this was found to be especially prevalent in surgical units.34 Another group found that patients readmitted within 30 days of hospital discharge had a high comorbidity index, frequent hospitalization in the last year, and a length of stay of five days or more.14,35 Readmitted patients may suffer from an unresolved condition, but they are also at a risk for other hospital-acquired issues that may arise such as opportunistic infections; therefore, prevention has become an important target for management.  The risk of unplanned hospital readmission is strongly influenced by SES.36–38 Many vulnerable populations do not have access to equitable resources and are disproportionally impacted by gaps in the healthcare system. There is often an absence of consideration for patients’ SES in the calculation of hospital readmission rates. Socioeconomic conditions can influence hospital readmissions through community-level marginalization and vulnerability.39 Previous studies have found that those living in marginalized communities, suffering from poverty, living alone, or unemployed were at an increased risk of readmission to the hospital.16,40,41   Many prediction models have been designed to identify at-risk patients early in the admission process and develop care plans that cater to the patient’s unique needs. Table 1.1 outlines some commonly used prediction models for preventive and exploratory purposes across a variety of healthcare settings. Some models are designed to assess avoidable 30-day readmission risk. The 6  HOSPITAL tool for example, may be used to flag vulnerable patients and trigger a referral to community-support programs that may encourage better adherence to the post-discharge care plan.42   There are limitations to relying solely on predictive models in clinical practice and when designing transitional care protocols. These tools have a limited capacity to distinguish between patients who will and will not be readmitted. Another limitation is the lack of calibration in the population of interest. A third limitation is in missing variables in the derivation datasets, such as SES, social supports, and clinical nuances. A fourth limitation includes an inability to calculate the predictive scores until the day of discharge which limits the window for pre-discharge interventions that might reduce readmission risk. Predictive models tend to stop at the point of discharge and fail to account for the dynamic nature of the post-discharge period. Despite these drawbacks, holistic approaches that also consider caregivers, social determinants, and patient’s preferences, may supplement prediction models in a way that overcomes some of their shortcomings.        7  Table 1.1 Common Indices and Prediction Models Used to Predict Risk of Hospital Readmission  Index/Model Variables HOSPITAL Hemoglobin at discharge; discharge from an Oncology service; Sodium level at discharge; Procedure during the index admission (any ICD-9-CM-coded procedure); index Type of admission (nonelective vs elective); number of Admissions during the past 12 months; Length of stay42 LACE Length of stay; Acuity of the admission; Comorbidity of the patient (measured with the Charlson comorbidity index score); Emergency department use (measured as the number of visits in the six months before admission)8 LACE+ Length of stay; Acuity of the admission; Comorbidity of the patient (measured with the Charlson comorbidity index score); Emergency department use (measured as the number of visits in the six months before admission); Age; Sex; Teaching status of discharge institution; Number of urgent admissions in previous year; Number of elective admissions in the previous year; Case-mix group score; Number of days on alternative level of care status43,44  Some studies have found that SES is associated with hospital readmission rates.36 Thus, measures of neighbourhood-disadvantage are used to help improve the prediction of readmission risks among a heterogeneous population and across health authorities. The predicative power of indices can give policy makers valuable insight to better plan for social programs that meet the specific needs of the community. In the U.S. an Area Deprivation Index (ADI) is used to measure neighbourhood-disadvantage.45,46 A recent study looking to the effect of ADI on hospital readmissions found that area-level variables strongly influence health care outcomes independent of characteristics at the individual-level.47 Researchers in Canada have also been looking at ways to analyze and measure deprivation and marginalization using indices that are 8  derived geographically and are area-based. The most recent iteration of this work is the Canadian Index of Multiple Deprivation (CIMD) released in 2019 which uses Census microdata.48 Using postal codes, the CIMD evaluates deprivation from a cross-sectional and dynamic lens that encourages a deeper understanding of SES.49 However, areal measures of neighbourhood SES, such as the ADI and CIMD are limited by the ecologic fallacy.50 Individual measures of SES cannot be modified and are often ignored by clinicians.51 Thus, the patient’s postal code can give us a snapshot of the community in which they live, but this information is most effective when used to supplement other more modifiable characteristics when striving to reduce the risk of readmission.  1.2.3 Interventions No single intervention has been conclusively found to reduce hospital readmissions in a reliable and predictable way; however; promising approaches have been identified that are not only patient-centered but can also combine pre- and post-discharge planning in a way that can bridge the gaps in transition.52 Over two decades of evidence-based work by Naylor et al. has led to the integration of transitional care in mainstream practice in a way that improves patient outcomes as well as achieves resource savings.53 Transitional care refers to timely services that are complementary to primary care and facilitate continuity of care as patients transfer from one level of care or setting to another.54 Table 1.2 outlines some common strategies, many of which are key components of transitional care interventions.    9  Table 1.2 Examples of Common Pre- and Post- Discharge Strategies             The performance of an individual hospital or healthcare team alone does not account for the variation in readmission rates. A recent systematic review of 11 transitional care interventions found that a combination of pre-discharge and post-discharge interventions had the most positive impact on reducing hospital readmissions.55 The discharge process, patient-centered care, follow-up after leaving the hospital, and community-based strategies are all equality important in the patient journey home.56   1.3 Virtual Care and Mobile Health Virtual care is defined as “any interaction between patients and/or members of their circle of care, occurring remotely, using any forms of communication or information technologies, with the aim of facilitating or maximizing the quality and effectiveness of patient care.”57,58 There is a movement towards using virtual care to improve access, make care more equitable, democratize health information, and reduce costs.29 However, there is a disconnect between consumer demand and availability. In a 2018 survey, 69% of respondents indicated an interest in virtual visits if available, while only 10% reported having had a virtual visit in the past.59 Given the  Pre-discharge   Post-discharge Patient education Community care Involvement of caregivers Telephone follow-up Interdisciplinary discharge planning Home visit Scheduling of primary care appointment Patient hotline Medication review Educational resources Community program referrals mHealth platforms 10  COVID-19 pandemic, demand for virtual care is expected to increase even higher.60 Advances in technology have created opportunities for creative approaches to healthcare with significant implications on assisting and improving continuity of care post-discharge.   The World Health Organization (WHO) defines mobile health (mHealth) as “the use of mobile and wireless technologies to support health objectives.”61 The growth of virtual care platforms are creating innovative opportunities for mHealth initiatives to integrate in patient care, both in and out of the hospital setting. Consequently, researchers advocate for user-centered, secure, and private technology that is accessible to the target population.62 When deploying mHealth initiatives, there is a potential for an intervention to further exacerbate marginalization. mHealth interventions that depend on smartphone ownership or cellular data connectivity may be inaccessible to patients with lower incomes, rural residence, or limited computer literacy. A 2017 estimate found that 31% of low-income households lacked internet access at home.63 Rural Canadians may also have limited internet and cellular data access. Thus, policy makers should be cognizant of variability in patients’ access, comfort, and preferences in using healthcare technology.29 One reliable way to overcome this gap has been the use of text-message based virtual care platforms.64 These platforms do not rely on applications to be downloaded or expensive data packages for internet access.   Although more patients have access to mobile phones, little is known about their usage preferences or limitations such as internet access, which may be restricted, particularly for marginalized populations for whom data plans are unaffordable.65 As more mHealth interventions are being deployed in clinical settings it is helpful to capture and assess patients’ 11  usage and preferences to ensure equitable access to health care resources.66,67 To this end, patients’ mobile phone use patterns and their sentiments towards mHealth interventions remain inadequately researched.68,69  1.3.1 WelTel Enthusiasm surrounding mHealth use has increased along with the rapid adoption of interventions aimed at improving patient care. Of the many emerging platforms, the majority utilize smart-phone applications that can be expensive and rely on a stable internet connection. WelTel is a pioneering virtual care mHealth platform that uses open two-way text messaging to enable patients to communicate with their care provider via mobile phone short message service (SMS). It allows the care provider to have access to an easy-to-use clinical dashboard where they can receive and send messages to patients in a secure and confidential manner. Furthermore, patients do not require access to the internet as messages are sent via SMS giving patients the ability to connect to their provider on their own terms and as needed.   This comprehensive program stemmed from work completed during the HIV epidemic in Africa where randomized control trials, by this thesis’ supervisor Dr. Richard Lester and colleagues, demonstrated improved adherence to medication using weekly check-in text messages to HIV positive patients.70 Since then, WelTel has expanded globally to include care in various settings and for diverse patient needs such as tuberculosis (TB), maternal and child health, and primary care.71–74 Although patient and provider satisfaction have been high in ambulatory settings, the platform has not yet been used in post-discharge and transitional care.75 Given the accessible and user-friendly nature of this platform, it is well-placed as a tool to support coordination of care as 12  it provides a secure and relatively inexpensive line of communication between discharged patients and their Health Care Provider (HCP). Throughout the journey of care, patients often navigate a long and complicated care path during which they may benefit from reliable communication with a trusted HCP.76 WelTel uses an “Ask, Don’t Tell” approach when communicating with patients. A weekly check-in message (e.g., how are you?) is sent to patients which prompts reflection and self-assessment. Furthermore, the open natural language communication can be used in any language and analyzed to examine the underlying causes of unplanned hospital readmissions.   1.3.2 Natural Language Processing and Machine Learning Rapid enhancements in health information technology have led to the availability of a large amount of structured (i.e., length of stay, age, and diagnoses) and unstructured (i.e., physician notes and narrative text) clinical health information data that is accessible via electronic health and medical records.77 Although this information is useful, it can be time consuming to sort through and expensive to analyze. Natural Language Processing (NLP) is a type of artificial intelligence technique that manipulates, interprets, and transports raw natural language text into information that is interoperable with computers.78 Machine Learning uses statistical methods and algorithms that enable computers to learn from data and make inferences with minimal human guidance.79 Together with emerging mHealth platforms, NLP and machine learning algorithms can be used to analyze patient conversations with healthcare professionals in an effort to understand emerging themes, gaps in care, and areas of concerns with regards to the coordination of care involved post-discharge from the hospital.  13  1.3.2.1 ConVIScope Conversational data analysis is a potential solution to the existing informational gap about patients’ needs after they are discharged from the hospital. Presently, one of the limitations to conversational data analysis is the time-consuming manual sorting of qualitative data such as physician notes or SMS. ConVIScope is a novel medical conversation analytics and visualization tool that uses NLP to examine patient-provider text message conversations in an environment that is secure and interactive. Data from previous deployments are used to train the underlying Machine Learning models which are supervised by subject matter experts. Using ConVIScope, care providers can examine the frequency of conversations associated with a topic of interest which can provide valuable insight to make important predictions (e.g., start of the flu season or mental health concerns among youth).80,81 Conversations can then be categorized based on topic, sentiment, urgency, or themes. Analyzing the patient’s conversation with their healthcare team might provide a better understanding of the underlying reasons why certain individuals are readmitted to the hospital soon after discharge.  1.4 WelTel Within the Context of The Patient’s Medical Home In 2011, the College of Family Physicians of Canada (CFPC) created the Patient’s Medical Home (PMH); a unified vision based on ten pillars, depicted in Figure 1.1, that built the foundations of a “one-stop-shop” where many of patients’ needs can be met.82 WelTel might complement some of the proposed visions for a comprehensive and patient-centered practice. As more primary care settings strive to implement the PMH, health care professionals will seek out partnerships with hospitals as many patients with chronic conditions find themselves caught in a cycle of readmission from which they often struggle to break free.17 The PMH emphasizes 14  patient-centeredness, community adaptiveness, and interprofessional collaboration. A practice that aligns with PMH can provide seamless, integrated, and timely care that can be a central hub for patients and their families. Although communities and clinics vary in their ability to adopt the PMH, technological innovations have evolved in a way that can enable many healthcare gaps to be bridged. Table 1.3 lists the specific attributes of each pillar within the core foundations of the PMH (bolded text highlight PMH elements that fit well within the WelTel mHealth platform).    Figure 1.1 The Patient’s Medical Home   Note: With permission. College of Family Physicians of Canada. A new vision for Canada: Family Practice—The Patient’s Medical Home 2019. Mississauga, ON: College of Family Physicians of Canada; 2019 15   Table 1.3 Patients Medical Home82 Theme Pillar Attributes      Foundations  Administrative Funding Governance, administrative, and management roles and responsibilities defined and supported; sufficient funding available; blended remuneration models.  Appropriate Infrastructure Use regionally approved EMRs and can access supports to maintain EMR systems; support collaboration and interaction; technology to enable alternative forms of care e.g., virtual care/telecare.  Connected Care Connection with community health and social services for timely patient referrals; continuity of patient information.            Functions  Accessible Care Access to medical advice, and information 24 hours a day, 7 days a week, 365 days a year; patient participation in planning and evaluation of appointment booking system.   Community Adaptiveness and Social Accountability Assess and address social determinants of health; use data about marginalized/at-risk populations to tailor care, programming, and advocacy; provide anti-oppressive and culturally safe care.    Comprehensive Team-Based Care with Family Physician Leadership Personal family physician and nurse form the core with roles of others (including but not limited to physician assistants, pharmacists, psychologists, social workers, physiotherapists, occupational therapists, dietitians, and chiropractors); co-locate team members or function as part of virtual networks.   Continuity of Care Foster long-term relationships between patients and care team; continuity of care in different settings e.g., hospitals, long-term care.  Patient- and Family-Partnered Care Active participation of patients, their families, and their personal caregivers in decision making process; encourage self-managed care; access to electronic medical records.     Ongoing Development Measurement, Continuous Quality Improvement, and Research  Establish and support continuous quality improvement programs; encourage and support community-based research. Training, Education, and Continuing Professional Development  Provide training environment for family medicine residents; enable sharing of knowledge with broader community; enable continuing professional education.   16  1.5 Research Questions and Objectives As highlighted above, unplanned hospital readmissions remain a challenge for patients, caregivers, clinicians, and policymakers. SMS-based mHealth interventions are a promising tool that might prevent unplanned hospital readmission. To assess the feasibility and effectiveness of this potential solution, we sought to address the following research questions:  1. Is mHealth accessible for patients admitted to the Clinical Teaching Unit (CTU) at Vancouver General Hospital (VGH)? a. To describe the mobile phone access, use, and preferences of CTU patients. b. To identify CTU patient characteristics associated with mobile phone access, use, and preferences.  2. In adult patients admitted to VGH CTU, is a two-way text messaging virtual care platform (WelTel) compared with regular post-discharge follow-up, effective in improving transitional care over the first 30 days following discharge?  a. To create a protocol for the design phase of the H2home WelTel pilot intervention.  b. To design a framework with which the underlying reasons for readmission to the hospital 30 days after discharge can be explored.  3. How can natural language processing (NLP) with machine learning (ConVIScope) be used to analyze two-way text message conversations between patients and health care providers (HCPs)? 17  Chapter 2: H2home Patient Access Survey  2.1 Background Over the last 25 years, Canadians have transitioned from landline service to widespread mobile phone use. Wireless technology is the fastest-growing sector in telecommunications with over 34 million subscribers in 2019.83 With the emergence of mHealth, mobile phones have become an important health care tool with which patients can have increased access to timely care. Innovative mHealth interventions are being sought in a variety of clinical settings and for a diverse range of chronic conditions.64,75,84,85 The versatility of these platforms suggests that they may be a cost-effective and promising way to improve transitional care and self-management support for patients.  Unplanned hospital readmission is costly and associated with adverse health outcomes; researchers are turning to mHealth technology as a possible solution.86 However, patients’ readiness, preferences, and access to mHealth platforms are infrequently examined prior to use. This is essential given regional and demographic disparities in access to technology. For instance, Canadians who live in urban areas have better access to internet via LTE on their mobile phones than those living in rural communities and phone plans with internet access can cost between $45 to $75, not including the cost of purchasing a cellular device.83 Completing an assessment and gathering sufficient data about the target population, setting, and stakeholders is critical because it allows program designers to address facilitators and barriers to implementation.87–89 To determine the accessibility, usage, and preferences of medical patients admitted to the Vancouver General Hospital (VGH) Clinical Teaching Unit (CTU) prior to the 18  deployment of a two-way text messaging mHealth pilot intervention, we conducted a survey to explore patient characteristics, mobile phone ownership, and communication preferences.  2.2 Methods We administered a cross-sectional survey of patients admitted to the VGH CTUs on January 7th, 2020, and January 23rd, 2020: a two-week period between survey dates was allotted to decrease chances of surveying the same pool of patients. Participants already surveyed on January 7th, and remaining in the hospital on January 23rd, were excluded the second time to avoid duplicate responses. The survey form was built on and data was collected using Qualtrics software, version January 2020.   The study took place in the VGH CTUs, which consists of 114 beds and provides acute medical care. Our inclusion criteria consisted of patients who were admitted to one of five CTUs, able and willing to provide informed consent, and able to complete the survey in English or via the aid of a proxy (i.e., spouse or child). We excluded i) patients residing in Long-Term Care Facilities (LTCF), as such patients are not dependent on mHealth given care received at LTCF, and ii) patients with significant limitations in their ability to interact with study team as per the discretion of the unit staff.  Each unit is assigned Care Management Leads (CMLs), consisting of nursing staff, who follow a discharge-planning protocol that entails the assignment of a Readmission Risk Assessment Score (RRAS) to each patient. The RRAS is based on a hospital prediction model, called the LACE index, and is modified to meet local needs (Appendix A). Prior to discharge, patients receive a 19  ‘Low’, ‘Moderate’, or ‘High’ RRAS depending on medical risk (such as an exacerbation of a chronic disease) and/or social risk (such as an inability to carry-out self-management activities). The CMLs provided a patient list of admitted individuals to the research team from which eligible participants were screened. Research staff approached eligible patients in their rooms, to provide information about the survey and gauge patients’ interest in participation. Patients who were interested were given information about the study and asked to provide informed consent. The research staff administered the survey orally and recorded the answers on a tablet computer. Patients who were sleeping or unavailable were re-approached later to ensure all eligible patients were given the option to participate. Participant’s sex, year of birth, RRAS, disease diagnosis, and the first three letters of the postal code were gathered and de-identified using the assignment of a unique patient identifier.  All statistical analysis was conducted using R 4.0.4. Descriptive statistics were used to characterize participants and summarize their mobile phone access, preferences, and usage. Inferential statistics were used to examine statistical significance. When analyzing nominal variables, Pearson’s chi-squared test was not performed as most variables had cell counts with an expected frequency of less than five which does not meet the assumptions required for a reliable relationship. Fisher’s exact test was used instead, to identify statistically significant associations for which contingency tables were analyzed. Next, Mann-Whitney U test was used to compare differences between ordinal or continuous and nominal variables. The Somers' Delta (Somers' D) and Goodman and Kruskal's gamma tests were performed for ordinal categorical variables to identify and measure the strength of any existing associations.   20  Table 2.1 Variables Considered and Method of Assessment to Conduct Statistical Tests Variables Assessment Sex Dichotomous nominal variable (male or female) Age Quantitative variable. Age groups Ages grouped by decade. Ordinal categorical variable. Disease Categories Disease diagnosis converted to disease categories. Discrete qualitative variable. RRAS Ordinal categorical variable (high, moderate, low) Phone ownership Dichotomous nominal variable (yes or no) Type of phone Dichotomous nominal variable (smart phone or other)  SMS use Dichotomous nominal variable (yes or no) Internet access Dichotomous nominal variable (yes or no) Most used Method Text Ranking of text message as most used communication method compared with voice and video. Ordinal categorical variable with three options (1, 2, 3) Ever texted with HCP Dichotomous variable (yes or no) Opportunity to text HCP Dichotomous variable (yes or no)   2.3 Research Question 1. Is mHealth accessible for patients admitted to the Clinical Teaching Units (CTUs) at Vancouver General Hospital (VGH)? a. To describe the mobile phone access, use, and preferences of CTU patients. b. To identify CTU patient characteristics associated with mobile phone access, use, and preferences.  2.4 Hypotheses H0i: There is no statistically significant difference in age by mobile phone ownership.  H1i: There is a statistically significant difference in age by mobile phone ownership. 21  H0ii: There is no statistically significant difference in sex by mobile phone ownership. H1ii: There is a statistically significant difference in sex by mobile phone ownership.  2.5 Results 2.5.1 Study Population The total number of CTU patients present on day one and two of the study were 107 and 113 respectively. Sixteen patients who were surveyed on day one, remained on the ward on day two and were therefore excluded from the survey on day two. Other exclusions included patients from LTCFs (n=27), patients who were non-responsive and/or study staff were given instruction not to approach (n=19), patients with a language barrier for whom no interpreter was present (n=16), patients from a corrections facility (n=2) given regulations regarding mobile phone ownership, and patients who were deceased (n=2). Most common reasons for non-response of eligible patients were “patient asleep” at the time of surveying (n=27) and reasons not captured (n=12). Therefore, 58.7% of eligible patients were successfully surveyed. See Figure 2.1 for participant flowchart.   Table 2.2 outlines descriptive information regarding survey participants. Respondents had a median age of 70 years (ranging from 30 to 98 years) and approximately 56% were male. Most participants were at a moderate risk of readmission as determined by their Risk of Readmission Assessment Score (RRAS). There were over 20 different disease diagnoses among the participants. Each diagnosis was associated with a disease category to facilitate a better understanding of the most common conditions. Figure 2.2 displays the disease categories upon admission to the unit. 22                         Day 1 Day 2 10C 23                                          10H 23 11A 2211D 24 14G 26 10H 23 10C 23 11A 23 11D 23 14G 25 118 117 Patients Screened 235 Eligible Patients 138 Participants 81 Exclusions: Long-term facilities, corrections, as per CML instruction, language barrier, deceased Non responders: Asleep, refused to participate, not in the room, reason undisclosed Figure 2.1 Participant Flowchart 23  Table 2.2 Survey Participants’ Characteristics Item Number Total Participants 81 Age in years (n=81) Median 70 Range 30 to 98  Frequency Percent Sex (n=81) Male 45 55.6% Female 36 44.4% RRAS (n=75) High 22 27.2% Moderate 36 44.4 % Low 17 21.0% Missing 6 7.4%   Figure 2.2 Most Responsible Diagnosis by Category   024681012141618Disease Category Number of Participants 24  Figure 2.3 Participants’ Neighbourhood of Residence Using First Three Letters of Postal Code Note: Map does not include those residing outside of the lower mainland. Points on map do not represent equal contribution from patients.  The most prevalent disease categories included infection (e.g., pneumonia), gastrointestinal (e.g., gastrointestinal bleed), cardiovascular (e.g., heart failure), and metabolic conditions (e.g., hyponatremia). Participants were from neighbourhoods across the lower mainland with the majority being within the Vancouver Health Service Delivery Area.90 Most participants resided in the city of Vancouver, however, there were also participants from Delta, Richmond, Langley, Surrey, Burnaby, North, and West Vancouver. (Figure 2.1). There were also three participants who did not reside in the lower mainland at the time of participation in this study. 25  2.5.2 Survey Responses The mHealth platform we intend to pilot uses a two-way short message service (SMS) to allow patients to communicate with their health care provider (HCP). To assess patients’ mobile phone access, we examined phone ownership and cellular features. Given that enrollment in mHealth may also be facilitated by a proxy (i.e., a trusted person with daily contact with the patient), we examined accessibility to a proxy for those without a mobile phone. Survey results are presented in Table 2.3 and Table 2.4, respectively.   Table 2.3 Participant Mobile Phone Access     Item Frequency Percentage Do you own a cell phone? (n=81) Yes 51 63.0% Proxy 19 23.5% No 11 13.6% If yes, is it a shared phone or a personal phone? (n=50) Shared 5 10.0% Personal 45 90.0% What type of phone do you use? (n=49)   Basic Phone (text/call) 8 16.3% Smartphone 41 6.1% Smart Phone (all types) 21 83.7% Does it come with a plan where you can call/text? (n=50) Yes, text and call 46 92.0% Call only 4 8.0% Text only 0 0% No 0 0% Do you have internet access on your phone? (n=48) Data + Wi-Fi 26 51.2% Wi-Fi only 10 20.8% None (call/SMS only) 12 25.0% 26  Table 2.4 Participant Mobile Phone Use   Item Frequency Percentage Do you text message on your phone? (n=49) Yes 37 75.5% No 12 24.5% Have you ever texted with your health care provider (e.g., doctor or nurse)? (n=73) Yes 18 24.7% No 55 75.3% Would you like the opportunity to text with your HCP? (n=74) Yes 53 71.6% No 21 28.4%   Table 2.5 Participant Ranking of Communication Methods by Frequency of Usage   Text Voice Video Ranked 1st 19 37.3% 32 62.8% 0 0.0% Ranked 2nd  16 31.4% 19 37.3% 16 31.4% Ranked 3rd 16 31.4% 0 0.00% 35 68.6%  We found that 63% of participants owned a mobile phone and 24% had access to a mobile phone via their proxy. Although most participants owned a smart phone, a quarter did not have internet access on their device. Notably, 75% of participants use SMS but less than 25% have ever texted their HCP. If given the opportunity, approximately 72% indicated an interest in text messaging their HCP. 90% of participants with access to a mobile phone, indicated that they can text and call using their device. Participants were asked to rank three communication methods in terms of usage frequency (text, voice, and video). Table 2.5 indicates that most participants chose voice as their first and text as their second choice. Interestingly, no one chose video as their first choice.    27  2.5.3 Accessibility and Usage Understanding participants’ accessibility and learning more about their usage is important for the development of a user-centered and engaging intervention. Table 2.6 shows the Fisher Exact Test evaluating the null hypothesis that there is no association between sex and technology use. Moreover, to verify the expected association between various mobile phone usages. Although no significant associations with sex were found, the expected associations between mobile phone usages are observed.   Confident levels of 95% or above are marked by an asterisk. Statistically significant relationships were identified and further analyzed using contingency tables as displayed in Appendix C (all percentages are calculated by column). The Bonferroni method is traditionally used to account for the multiple hypothesis problem; however, it is a conservative measure that can result in many false negatives.91 The false discovery rate (FDR), proposed by Benjamini and Hochberg, is the expected proportion of incorrect rejections between all rejections.92 I calculated the FDR to identify as many statistically significant associations as possible while maintaining a low false positive rate. Our FDR is 15%; in other words, two of the rejected null hypotheses may be due to sampling variability or by accident.      28  Table 2.6 Correlation matrix between variables using Fisher’s exact test   Sex Phone Ownership Proxy Type of Phone Internet Access SMS Use Ever Text HCP Like to Text HCP Sex - 0.12 1 0.18 0.17 0.5 0.27 0.31 Phone Ownership 0.108 - 0.387 NA NA NA 0.258 0.414 Proxy 1 0.387 - NA NA NA 0.205 0.001* Type of Phone 0.178 NA NA - 7.65E-6* 0.001* 0.416 0.018* Internet Access 0.169 NA NA 7.65E-6* - 2.32E-5* 1 0.049* SMS use 0.494 NA NA 0.001* 2.32E-5* - 0.414 0.136 Ever Text HCP 0.272 0.257 0.205 0.416 1 0.414 - 0.028* Like to Text HCP 0.310 0.414 0.001* 0.018* 0.049* 0.136 0.028* - *P<0.05; FDR = 0.15; NA = Not applicable (each variable has a column total equaling zero) 2.5.4 Age and Sex To determine if there were associations in the variables and sex, Fisher’s exact tests were performed. Phone ownership and sex resulted in a p-value of 0.12. Thus, we fail to reject the null hypothesis and conclude that our sample does not provide sufficient evidence of an association between sex and phone ownership.  To examine whether age differs by phone ownership status, Mann-Whitney U tests were performed. We found a statistically significant difference, with a p-value of 0.0008, in mobile phone ownership by age (Figure 2.4). Thus, we reject the null hypothesis and conclude that there is evidence of a statistically significant difference in age by phone ownership. We also found a statistically significant difference in age by internet access (Figure 2.5) and texting messaging (Figure 2.6). The average age of participants who did not own a phone, have mobile internet access, and did not text message was over the age of 70.  29  p-value = 0.039 p-value = 0.0009 p-value = 0.0008 Figure 2.4 Age of Mobile Phone Owners Figure 2.5 Age of Patients with Internet Access Figure 2.6 Age of Patients Using SMS                        30  2.5.5 Risk of Readmission Assessment Score (RRAS) To further examine participant data and survey responses, several explorative analyses were performed. First, the Risk of Readmission Assessment Score (RRAS) and age were analyzed using the Somers' Delta (Somers' D) and Goodman and Kruskal's gamma tests, however no statistically significant associations were found. Next, a Mann-Whitney U test was performed for some ordinal and nominal variables to explore any associations. Contingency tables for these results are portrayed in Tables 2.7 and 2.8 respectively.  Table 2.7 One-way Texting with HCP and Risk of Readmission Assessment Score (RRAS) RRAS Would Like Opportunity to Text HCP to Receive Health Information (one-way) Total No Yes Low 12 32.4% 5 15.6% 17 Moderate 18 48.6% 14 43.8% 32 High 7 18.9% 13 40.6% 20 Total 37 100.0% 32 100.0% 69 p-value = 0.03 Table 2.8 Two-way Texting with HCP and Risk of Readmission Assessment Score (RRAS) RRAS Would Like Opportunity to Text HCP to Discuss Healthcare Concerns (two-way) Total No Yes Low 11 33.3% 6 16.7% 17 Moderate 16 48.5% 16 44.4% 32 High 6 18.2% 14 38.9% 20 Total 33 100.0% 36 100.0% 69 p-value = 0.03 31  We found a statistically significant difference in participants’ interest in texting their HCP to receive health information (one-way communication) and to discuss their health (two-way communication) by their RRAS. To clearly identify the direction and magnitude of interest in texting HCPs, we calculated the prevalence ratios of interest in one-way and two-way texting among those with a high and low RRAS with 95% confidence intervals in Table 2.9. In the numerator we have the number of patients interesting in texting their HCP. In the denominator we have the total number of patients with high RRAS and total number of patients with low RRAS.  Table 2.9 Measures of Association for Two-way and One-way Communication with HCP  Proportions Group of patients with high RRAS  Group of patients with low RRAS Prevalence ratio (95% CI) Would Like Opportunity to Text HCP to Discuss Healthcare Concerns (two-way) 14/20 (70%) 6/17 (35%) 1.98 (.98, 4.01) Would Like Opportunity to Text HCP to Receive Health Information (one-way) 13/20 (65%) 5/17 (29%) 2.21 (0.99, 4.94) 32  While 70% of participants with high RRAS would like the opportunity to text HCPs to discuss healthcare concerns, 35% of those with low RRAS would like the opportunity to text HCPs to discuss healthcare concerns. This shows a prevalence ratio of 1.98 with 95% confidence interval between .98 and 4.01. While 65% of participants with high RRAS would like the opportunity to text HCPs to receive health information, only 29% of those with low RRAS would like opportunity to text HCPs to receive health information. This shows a prevalence ratio of 2.21 with 95% confidence interval between 0.99 and 4.94. In other words, the proportion of participants interested in two-way communication and one-way communication with their HCPs is approximately 2-fold greater if a person is at a high risk of readmission. Patients at higher risk of readmission were more likely to want the opportunity to receive one-way text messages and were more likely to want the opportunity to discuss health care concerns with HCPs.  2.6 Discussion The survey results presented show that access to mobile phones is not the same or equal among the patients admitted to the CTUs at VGH. Evidence that there is unequal mobile phone access is that some patients do not own a mobile phone, or they must rely on a proxy for access. Of those who depend on their proxy for mobile phone access, 87.5%. would like the opportunity to text their HCP. In fact, age was an important factor that impacts accessibility. Evidence of an age disparity is that 80% of those without a mobile phone were over the age of 60 (Figure 2.4).  The mobile phone usability and preference data obtained in this study suggest that SMS-based interventions might have better up-take in this patient population and programs that rely on smart-phone features and internet connectivity would be less accessible for certain patients. This 33  has been demonstrated in three ways. First, we found that a quarter of participants with a mobile phone did not have internet access on their device. Interventions that require connectivity to the internet would be inaccessible to these patients. Of those without internet access, 25% indicated that they use SMS and 50% would like the opportunity to text their HCP. Second, we found that 21% of participants did not have access to a smart phone. Of those without a smart phone, approximately 82% did not have internet access. Of those with a smart phone, 89% indicated that they use SMS and 84% would like the opportunity to text their HCP. Third, approximately 72% of participants would like the opportunity to communicate with their HCP via SMS and of those who have never texted an HCP, 65.5% would like the opportunity to text an HCP.   In recent years, many medical technology developers have focused on smart phone applications as a mode of service delivery. However, our data show that not all medical participants have access to the internet on their mobile devices, nor may know how to use ‘apps’. It is possible that patients who require the most of HCPs’ time and are at the highest risk of an unplanned hospital readmission, are also unable to use smartphone applications. In contrast, many participants can access a basic mobile device that can send and receive phone calls and text messages. Thus, the greatest gains in mHealth interventions may come from ensuring adequate reach to the target patient population.   There are several limitations that may introduce bias and reduce generalizability. First, our sample size was limited due to research activity cessation in response to the COVID-19 pandemic. Second, we did not screen patients with language barriers as it prevented them from understanding survey questions or providing informed consent. In future iterations of this study, 34  we aim to utilize interpretation services to better understand the access and phone preferences of this population as language barriers might also increase risk of readmission and other poor health outcomes. Third, although initial demographic comparisons show no significant differences between participant responders and non-responders in terms of sex and age, we expect some nonresponse bias. For example, patients unable to participate, might be of poorer health or have cognitive impairments that might also increase their risk of readmission. Finally, we surveyed patients in the CTUs, at a large academic teaching hospital, where the results may not be generalizable to other urban hospital inpatients. This study population was conveniently sampled which may have introduced biases through patient selection. This specific population was selected as they are part of an active and already funded project by the UBC mHealth Research Group. Future iterations of this survey should include a variety of inpatient hospital wards in an urban clinical setting.  These results show that most participants had access to a mobile phone either through ownership or via a proxy (e.g., family or caregiver). Thus, phone-based mHealth interventions are accessible and those that allow enrollment via a proxy, could improve inclusivity, and facilitate care-giver engagement. Furthermore, participants indicated significant interest in communicating with HCPs via text message; given that our planned intervention is text-messaged-based and does not rely on internet access, it is well placed to bridge a potential technological and equity gap for those who may not be able to access internet-based mHealth platforms.  35  Chapter 3: H2home mHealth Intervention Pilot Development 3.1 Introduction Vancouver Coastal Health (VCH) is committed to improving the patients’ transition from the hospital to home after discharge; mHealth presents an opportunity to fill existing gaps using innovative interventions. Hospital administrators often utilize quality improvement interventions to address clinical and psychosocial challenges that impact inpatient and outpatient care.93–95 In transitional care, high readmission rates signal gaps in continuity and coordination after discharge from the hospital. Likewise, from a health systems perspective, readmissions are an area of concern given its costs and negative impact on patients’ health. Among Canadian provinces, British Columbia (BC) has the second highest rate of readmission to the hospital within 30 days of discharge. VCH is one of four health authorities in BC covering over 1.2 million people. The 2019-2020 risk of readmission to the hospital was estimated at 10.0% for BC and 10.3% for VCH, both of which are higher than the national average of 9.5%.2   In response to high readmission rates, VCH implemented two care models in 2014 – iCare and the Ideal Transition Home (ITH).96,97 iCare is a care planning methodology aimed at the delivery of quality care, reducing delays, and eliminating barrier to a safe discharge from the hospital. ITH is a proactive initiative that identifies patients at risk for readmission to facilitate a safe and timely transition home. iCare and ITH are in line with VCH’s “Home is Best” philosophy and were developed and implemented by an interprofessional care team; not only have these programs improved the patient’s journey of care, but they also reduce patient’s length of stay and free hospital beds.98  36  Within 48 hours of admission to the hospital, the team begin developing a picture of the patient’s clinical needs and barriers to returning home. The care team’s goal is to develop the best post-discharge plan and to create early linkages with community partners; thus, the patients who are at a risk for returning to hospital following discharge are identified. At this time, the family physician of the patients will be informed of the admission to the hospital with the goal of scheduling an appointment within 48 to 72 hours of the patient returning home. There is increased focused on a collaborative relationship between the hospital and community partners. In preparation of the patient’s discharge, pertinent resources such as the ‘My Discharge Plan’ (Appendix B) will be shared with the patient. This information is in easy-to-understand language and serves as a supplement to patient education about the signs and symptoms of their condition.   iCare and ITH ensures consistency and continuity of care by following-up with the family physician and other relevant healthcare partners in the community while taking into consideration the patients’ preferences and place of residence. The iCare program empowers acute care providers to be in sync with primary care while also respecting the patient’s preferences and to give patients an understanding of their illness which can lead to the prevention of further deterioration, emergency visits, and subsequent hospital admissions.96,98 However, hospital readmission rates are still an ongoing problem particularly for patients with mental illness, substance use disorders, and other comorbidities.99 A first step in understanding the multifaceted causes of hospital readmissions is understanding its cyclical nature (Figure 3.1). Policy makers have experimented with multiple interventions ranging from incentives to penalizations.100 In the context of the Canadian health care system, approaches that utilize technology, such as mHealth, to overcome gaps in care can have great potential. 37  Figure 3.1 Hospital Discharge and Readmission Cycle   Most studies on hospital readmission focus on medical chart reviews or medical claims data, thus, there is a need to fill an existing knowledge gap by augmenting our current knowledge with patient experiences. Although, patient surveys have been used in the past to gain a better understanding of the reasons for readmission, they are not as reliable given the risk for courtesy or acquiescence response bias.17   In the proposed study we will utilize WelTel, a two-way text messaging virtual care service, that enables patients to communicate with their healthcare team after discharge from the hospital in open natural language. This service allows the healthcare team to follow-up with patients and provide streamlined support that is accessible and confidential. In the past, narrative responses, such as patient conversations, would have to be coded, analyzed, and placed within themes. This can be very labour intensive and time-consuming. We will use a health conversation analytics and visualization tool, called ConVIScope, to analyze the text messages and better understand Hospital AdmissionRisk of Readmission Assessment ScoreDischargePlanningPatient Discharged from HospitalReturn HomeEmergency38  the underlying reasons for hospital readmission. Examining themes among patient conversations with HCPs protects the data from being impacted by the patients desire to give positive responses that they think their HCPs or research staff would want to hear. Using this novel approach, we can create a learning health system to gain valuable insight from patient conversations, improve continuity of care, and provide a better transition from the hospital to the patient’s home.101  3.2 Thesis Objectives 1. To plan and develop a protocol for the WelTel H2home Intervention phase. 2. To develop a framework to understand the underlying reasons for unplanned readmission to the hospital. 3. To pre-test and train the ConVIScope conversation analytics software.  3.3 Research Aims 1. To determine if using the WelTel virtual care platform at VGH can help support and streamline transitional care and follow-up of from hospitalised patients on the CTU in the first 28-days after discharge.  2. To determine if using the WelTel virtual care platform improves patient experience and satisfaction with care post-discharge from VGH CTU. 3. To determine if unplanned hospital readmission rates can be reduced after introducing WelTel for six months among patients discharged from the CTUs.  39  3.4 Study Interruption Only weeks before the date of our study deployment, all research activity at the hospital was halted due to the global COVID-19 pandemic. We decided to shift our program to adapt to these changes as this study is well-placed to fill an existing and important gap in data during this pandemic. Given that COVID-19 positive patients and other patients at high risk may have unique follow-up requirements after discharge from the hospital, we proposed that our program would be an efficient way to support patients through their post-discharge journey. While we felt this project would be ideal, the hospital flow changed rapidly, and personnel were unable to engage in research activity as they addressed the pandemic priorities. We will resume the project once research activities are permitted to recommence at the hospital.   3.5 Methods Vancouver General Hospital (VGH) is a large teaching hospital located in British Columbia (BC). The VGH Clinical Teaching Units (CTUs) with 114 acute medical beds serves approximately 4,400 patients yearly. There are Care Management Leads (CMLs) assigned to the CTU who oversee the completion of the Readmission Risk Assessment Score (Appendix A) and “My Discharge Plan” (Appendix B). The mHealth platform used in this intervention is called WelTel H2home. Given that WelTel is a configurable platform, it was adapted for this project in consultation with the CLMs to automate some components of post-discharge follow-up and transitional care. WelTel H2home provides a secure dashboard where a health care provider (HCP) can send messages to and receive text messages from patients’ mobile phones. Furthermore, the platform will enable patients to communicate with their HCP using any mobile device.  40  Table 3.1 WelTel H2home Features Feature Components Patient registration Using any mobile phone. Weekly automated message How are you? Removal from the study  Can be done verbally with a healthcare provider or by texting “STOP” to the WelTel phone number. Appointment reminders e.g., blood test and immunizations. Virtual appointments e.g., real time chat, phone, and video. Patient Portal Digital version of “My Discharge Plan, Accessible via link sent by text message or by pasting link into internet. Provider Dashboard Flags patient messages that need attention, allows HCPs to respond.  Study participants must be able to read and understand English to provide informed consent. We will not be utilizing translation services because the WelTel platform is currently only offered in English and French. Furthermore, patients will be communicating with the HCPs, most of whom only speak English, in an ad hoc way making it difficult to utilize translators or interpreters. To address this issue, we allow individuals to enrol via a proxy who can communicate in English and with whom participants have near daily contact.   To be eligible to participate in this study, patients will need to a) be admitted to one of the five CTUs participating in the study (10C, 10H, 11A, 11D, or 14); b) either have a mobile phone with a Canadian number, or a proxy (e.g., family member or care giver with near daily contact) with a mobile phone with a Canadian phone number; c) be deemed appropriate by the CML; and d) provide informed consent. Patients who do not meet the above criteria will be excluded from the intervention but will still receive the regular standard of care offered to patients discharged from the CTU. A pseudo-randomization approach will be used by randomly assigning one unit per floor to be an intervention unit. The other unit will act as a comparison group (i.e., 10C = 41  intervention, 10H = control). An exception is CTU 14G, which will be assigned as an intervention unit in its entirety due to the lack of an appropriate comparison group. VCH’s Decision Support unit will aid in the analysis and access to discharge data which are collected in the administrative databases. To ensure anonymity and prevent any risks to patient privacy, hospital medical record numbers (MRNs) will be matched to study IDs and a crosswalk file will be kept in a separate location for linkage purposes only. The WelTel enrolment, usage, and communication data will be saved in the WelTel platform and exported for analysis.          A link to a patient satisfaction survey will be sent via text-message to all participates at the end of the 30-day follow-up period. If participants do not respond to the survey within seven days, a research coordinator will phone the participant and provide the option of completing the survey in written form or by phone call to enhance data capture. One focus group discussion (FGD) will be held at the beginning of the active pilot and another after the 6-month pilot period. The FGDs will invite 5-10 CMLs, CTU clinical team members, and patients to discuss their experiences using a modified Consolidated Framework for Implementation Research (mCFIR) tool. We hope to gain a better understanding of the CMLs attitudes towards current discharge planning procedures and recommendations for how these programs can be improved.  Figure 3.2 H2home Intervention Pilot 42  3.5.1 Patient Consent Once patients are admitted to VGH and assigned to a unit, they will be associated with a CML team. This team begins their discharge planning procedures as soon as the patient is admitted to the unit and ensures that all the necessary paperwork is complete prior to the patient being discharged from the hospital. All patients admitted to one of the five CTUs, who meet the inclusion criteria, will be informed of the intervention by the CML. Patients who express an interest in learning more about the study will be flagged and referred to the study coordinator. The study coordinator will approach the patients in their room to review the study procedures and obtain informed consent. The consenting process will take between 20 minutes up to an hour, depending on the person and their unique circumstances. Due to the current COVID-19 pandemic, patients who test positive for COVID-19 will be consented via telephone as per UBC guidelines for obtaining virtual consent. Once enrolled in the WelTel H2home program, they will be able to use it as soon as they are discharged from the hospital. The amount of time spent using the platform and engaging with their HCPs via the platform, is entirely up to the participant and their individual health care needs, concerns, and care plan.   To ensure a consenting protocol that is fully voluntary and not coerced, participants will receive detailed information about their rights as a research participant and their ability to withdraw from the study at any time they wish without consequences to the care they are receiving. We will inform participants of the potentials for breach of confidentiality and the ways we plan to mitigate the risk. All study-related data will be identified only by a study identification number. Data will be maintained on an encrypted password protected computer and stored securely at the study site. Only designated individuals will have controlled access to study data. The cell phone 43  number used by the care provider will be a ‘private’ number not associated with a medical institution or research study. Participants will not incur any personal expenses because of participation and participants will not be paid for participating.   3.5.2 Patient Registration Patients admitted to the CTUs usually get screened within 48 hours. In addition to the standard of care typically received, patients who consent to participate in the study will be enrolled in the WelTel virtual care service adapted to support the discharge planning process. The core research-team members met regularly with CMLs to ensure stakeholder input as we designed the study protocol. We wanted to ensure patient recruitment and registration would not be disruptive to patient care and would not increase the workload of staff. Working collaboratively with unit staff allowed the research team to gain trust as well as pre-emptively anticipate and address problems. We conducted a training session with the CMLs as well as the nursing team who would be using the platform. We received feedback regarding the user-interface and changes were made accordingly.  Figures 3.3 to 3.5 are screenshots of the patient registration process. There are both optional and mandatory fields available to the staff during the initial registration of patients into the platform (Figure 3.3). The mandatory fields include the patient’s name (each patient will be assigned a unique ID upon data-export to de-identify all personal information), year of birth (to calculate age while respecting confidentiality), MRN number (to match readmission data received from the Decision Support Unit), primary mobile number (required to receive text messages and 44  utilize the WelTel service), and assigned clinic (this will be the unit to which the patient is admitted; either 10H, 10C, 11A, 11D, or 14G).            Figure 3.4 displays the next section of registration. Here, the proxy option will be clicked for participants without a mobile phone but with near daily access to a trusted caregiver or family member with a mobile phone. The participant will then be able to communicate with their HCPs using their proxy’s mobile phone. The next field provides the HCPs with the ability to adjust the check-in frequency, which can be set according to patients’ preferences and needs with the default being a weekly “how are you?” text message. The open-ended nature of this question encourages patient engagement and facilitates better communication with the HCPs; however, this question can be modified and personalized at the discretion of the CMLs.  Figure 3.3 Participant Registration (part 1) 45  Figure 3.4 Participant Registration (part 2)   Figure 3.5 outlines the final step in the patient’s registration. As soon as the participant’s Risk of Readmission Assessment Score (RRAS) is determined, it will be checked off in the ‘Groups’ box. The research consent box will be check-off by the research-team to ensure all participants have signed the informed consent form. Lastly, once the patient is discharged home and completes the post-study survey, this box will be checked off to signal completion.         Figure 3.5 Participant Registration (part 3) 46  3.6 Analysis Plan We will leverage the extensive knowledge and expertise of the VCH Decisions Support Unit team who normally report on and analyze data such as hospital readmissions and discharges. First, all data related to the study will be de-identified and only indicated by a study number. Second, the data will be kept on an encrypted and password protected computer at a secure study site located at VGH. Finally, only designated study staff will have access to study data and any documents related to the participants. We will compile and analyze the survey data in descriptive form using R software, and quantitative data from the FGDs will be transcribed and evaluated using NVivo software and Qualtrics. An interim analysis will be conducted at six months. All evaluations will be adjusted for patient-, provider-, and hospital level- covariates. Based on recent data, approximately 640 patients are discharged from the CTU each month. The expected enrolment is at approximately 80% which translates to about 768 patients enrolled over the 6-month pilot. Based on this projected enrolment, this study is expected to have an >90% power to detect an absolute risk reduction of 1% or more (alpha = 0.05; sigma 3.0; 2-tailed).  To support the analysis plan of WelTel H2Home, we will be using the ConVIScope software system as an add-on component to the WelTel platform to explore themes within conversations and to gain a better understanding of the circumstances and underlying reasons for hospital readmissions. ConVIScope, is a product of the mHealth Research Group in partnership with the Computer Science Department. ConVIScope uses a patient conversation analytics machine learning software which can cleanse and format any patient-to-provider conversational data sets, train A.I. models, classify patient voices into clinical topics, and a navigate large volume of conversational data in an interactive fashion. This novel technology can empower care teams to 47  improve quality of care by examining patient interactions, sentiments, and concerns. Our work is part of the development and optimization of ConVIScope.  3.7 Limitations Our study will have several strengths and limitations to mention. First, we will be using data from one hospital, for this reason any findings may not be generalizable to other hospitals and communities. Second, we will be capturing patient experiences via a survey which may introduce bias given that is self-report. However, receiving this feedback will give us insight into patients’ values and it is a key step in understanding patient preferences in this pilot. Third, given the nature of FGDs, it may be difficult for staff to be open regarding some of their opinions for fear of it negatively impacting their work. To mitigate this, we will ensure FGD facilitators are not associated with the hospital. Lastly, lack of fluency in English is a potential barrier and may decrease inclusivity. In future iterations of this intervention, we plan to utilize VGH patient services that provide sign language interpreters as well as spoken language interpreters.  3.8 H2home WelTel Framework In preparation for the re-launch of the pilot program, we developed a framework consisting of four main steps to understand the underlying reasons for readmission to the hospital (Figure 3.6).  The first step, completed in December 2020, consists of the pre-intervention phase to ensure sustainability and stakeholder engagement. Any usability issues were anticipated and addressed before proceeding to the subsequent steps. Steps 2 to 4 take place in a continuous and simultaneous manner in consultation with the primary investigator, decisions support unit, hospital staff, and relevant stakeholders.  48  Understanding Reasons for ReadmissionPatient Conversation DataSocioeconomic StatusClinical Risk FactorsPost-study SurveyFocus Group Discussion  • Check-in with discharged patient for 30-days • Monitor patients on WelTel Dashboard • Follow-up on patients’ or their family members’ concerns • Collaborate with interprofessional team • Communicate with community team Figure 3.6 H2home WelTel Hospital Readmission Prevention Framework            Assessment• Phone Access Survey• Stakeholder ConsultationsDesign• Study Protocol• Ethics ApplicationDeployment• CML Training• Stakeholder Engagement• Follow-up questions • New/worsening symptoms • Care-giver concerns • Test-results • Appointment reminders • Changes in Medication • Follow “My Discharge Plan” • Communicate with CML for 30 days • Attend follow-up appointments • Adhere to medication regiments • Complete post-study survey • Family able to communicate with CML  49  3.9 ConVIScope Training To understand and analyze patient conversations, which will be available to us in the form of unstructured text, we need to use artificial intelligence and a machine learning model. This model changes as it acquires more “knowledge” which it “learns” from training data. A first step in machine learning for natural language processing (NLP) is the supervised learning phase.102 We began our training by annotating patient conversations in our training data. We tagged these patient conversations using pre-determined categories that were defined by our clinical knowledge experts. The general topics include: Symptoms, Diagnostic, Treatment (Rx), Lifestyle, Social, Logistical, Health Education, Service Quality, Urgency, and Special Topics.   Figure 3.9 displays an example of a supervised learning session. In this example, I have annotated the conversation as “Procedure” and “Scheduling” because the patient is discussing an appointment during which the procedure of an IUD insertion will take place. Next, the patient is discussing being away which I have annotated as “Travel.” Finally, I have annotated this conversation as “non-urgent” because it pertains to an issue that could be addressed over weeks as opposed to days or hours. To date, I have annotated over 1200 text message conversations between patients and HCPs. Together, with other individuals on our research team, we meet regularly to ensure consistency in our annotations and refine the definitions as we prepare to use a new dataset to re-train the model.        50  Figure 3.7 Supervised Machine Learning of Patient Test Training Data    3.10 Next steps We hope that our findings will give hospitals the tools to work cohesively with primary care physicians and social workers in the community. Furthermore, an integrated mHealth platform can aid in the delivery of timely, accessible, and comprehensive care. Looking into the future, the findings of this study can also address the Patient’s Medical Home pillars outlined by the College of Family Physicians of Canada (CFPC). There are many benefits to aligning a practice with the PMH principles such as improved access to care, higher patient and provider satisfaction, lower ER visits and hospitalizations in patients with chronic disease, better quality of care, and improved cost savings.103,104 The proposed framework will also inform other virtual care opportunities and internal decision making at VGH. 51  Chapter 4: Conclusion There has been interest in hospital readmissions since as early as the 1970s but there was added interest when a high prevalence of rehospitalizations among Medicare beneficiaries was reported in the 1980s.105 While early studies generally focused on the geriatric population, policy makers have a lot to gain from reducing hospital readmission rates for all patient groups. Furthermore, readmissions are associated with patient dissatisfaction which has important implications on many aspects of patient care such as medication adherence, continuity of care, and trust in the medical system.5 Promising approaches to reducing hospital readmissions reflect the need to improve post discharge coordination of care, reduce healthcare spending, and increase effective interventions that focus on prevention. Although many studies have examined the clinical characteristics influencing readmissions, the underlying reasons for hospital readmissions and the unmet needs of patients are less known.  An increasing body of literature seeks to find ways of lowering hospital readmission rates around the world.18,35 Insight into this issue can lower costs, improve quality of care, and remove some of the burden on hospital staff. Innovative and sustainable interventions involve patients and HCPs in the assessment and evaluation of mHealth initiatives; programs can benefit from involving end-users at the design, creation, testing, and evaluation phase.62 Trust in mHealth services can impact trust in future mHealth interventions and even have implications on trust in the healthcare system. A 2020 survey of 1,800 Canadian residents found that even among those with a family doctor, 15% access the emergency department should an illness or medical condition arise; once the global pandemic is resolved, 38% would choose the option of phone, video conference, email, or text message rather than an in-person visit.106 Although the field of 52  virtual care has been growing rapidly, the incredible speed with which it was adopted to all aspects of life in response to the global COVID pandemic is remarkable.  Our survey study of patients admitted to the medical units at VGH CTU shows that there are differences in accessibility, usage, and preferences among participants. Notably, there are differences in phone ownership, internet access, and SMS usage by age. However, there were no statistically significant differences by sex. This may be due to our small sample size and should be examined in subsequent studies. By using a patient-centered approach that gives special consideration to older patients and those with a high risk of readmission, SMS-based mHealth technology is well-placed and accessible for this patient population.  We designed the H2home WelTel intervention to understand and to address high unplanned hospital readmission rates; we have outlined a framework (Figure 3.6) that can be used to implement this program in other healthcare settings. This mHealth intervention, which incorporates and compliments existing discharge planning and transitional care initiatives, is well-placed to support the province’s existing move towards a Patient’s Medical Home (PMH). Once implemented, the coordinated nature of the WelTel H2home intervention can create links between the primary care environment and the hospital system which is line with the PMH pillars.82     53  There are several novel aspects in this study that draw on previous research indicating a need for an integrative and collaborative approach to mHealth.61,75 First, the patient survey both before and after the study gives patients an opportunity to provide feedback at various points in the intervention cycle. Second, ConVIScope includes unique natural language processing and machine learning features that may provide a more complete understanding of the underlying reasons for unplanned hospital readmissions among medical patients. It can empower care teams to explore patient voices and to discover precision population health using natural language processing (NLP), machine learning, and data visualization technologies. Lastly, focus group discussions (FGDs) will allow stakeholders to engage with the research and provide real-time feedback.  We have shown that designing a patient-centered and equitable program is feasible using SMS-based mHealth technology in collaboration with two important stakeholders, namely HCPs and patients. In both cases there is adequate interest and accessibility to mobile phones to facilitate a low-cost and sustainable intervention. Furthermore, we have built a foundation upon which we can test natural language processing and machine learning to analyze patient conversations with their HCPs to gain a deeper and better understanding of underlying reasons for readmissions to the hospital. This information may facilitate lower hospital readmission rates by improving transitional care, patient engagement, and continuity of care.    54  References 1.  Canadian Institute for Health Information. All-Cause Readmission to Acute Care and Return to the Emergency Department Health System Performance. Heal Syst Perform. Published online 2012:64. 2.  Canadian Institute of Health Information. Your Health System: Patients Readmitted to Hospital.; 2020. Accessed December 30, 2020. https://yourhealthsystem.cihi.ca/hsp/inbrief.#!/indicators/006/all-patients-readmitted-to-hospital/;confidence;mapC1;mapLevel2;overview;provinceC5001;trend(C1,C9001);trendHospital(O80284);/ 3.  Staples JA, Thiruchelvam D, Redelmeier DA. Site of hospital readmission and mortality: a population-based retrospective cohort study. Can Med Assoc J Open. 2014;2(2):E77-E85. doi:10.9778/cmajo.20130053 4.  Shuster C, Hurlburt A, Tam P, Staples JA. Unplanned hospital readmissions in British Columbia. BC Med J. 2018;60(June). 5.  Boulding, William P, Glickman, Seth W., MD M, Manary, Matthew P. M, Schulman, Kevin, A. M, Staelin, Richard P. Relationship Between Patient Satisfaction With Inpatient Care. Am J Manag Care. 2011;17(1):41-48. 6.  Kansagara D, Englander H, Salanitro A  et al. Risk prediction models for hospital readmission: A systematic review. JAMA - J Am Med Assoc. 2011;306(15):1688-1698. http://jama.ama-assn.org/content/306/15/1688.full.pdf+html%5Cnhttp://ovidsp.ovid.com/ovidweb.cgi?T=JS&PAGE=reference&D=emed10&NEWS=N&AN=2011580014 7.  Joynt KE, Jha AK. Who has higher readmission rates for heart failure, and why? 55  Implications for efforts to improve care using financial incentives. Circ Cardiovasc Qual Outcomes. 2011;4(1):53-59. doi:10.1161/CIRCOUTCOMES.110.950964 8.  Van Walraven C, Dhalla IA, Bell C, et al. Derivation and validation of an index to predict early death or unplanned readmission after discharge from hospital to the community. Cmaj. 2010;182(6):551-557. doi:10.1503/cmaj.091117 9.  Clarke A. Are readmissions avoidable? BMJ. 1990;301(6761):1136-1138. doi:10.1136/bmj.301.6761.1136 10.  Van Walraven C, Bennett C, Jennings A, Austin PC, Forster AJ. Proportion of hospital readmissions deemed avoidable: A systematic review. CMAJ. 2011;183(7). doi:10.1503/cmaj.101860 11.  Shuster C, Hurlburt A, Yung T, Wan T, Staples JA, Tam P. Preventability of 28-Day Hospital Readmissions in General Internal Medicine Patients: A Retrospective Analysis at a Quaternary Hospital. Qual Manag Health Care. 2018;27(3):151-156. doi:10.1097/QMH.0000000000000174 12.  All Patients Readmitted to Hospital · CIHI. Accessed April 21, 2020. https://yourhealthsystem.cihi.ca/hsp/inbrief?lang=en#!/indicators/006/returning-to-hospital/;mapC1;mapLevel2;/ 13.  All Patients Readmitted to Hospital · CIHI. Accessed April 4, 2021. https://yourhealthsystem.cihi.ca/hsp/inbrief.#!/indicators/006/all-patients-readmitted-to-hospital/;mapC1;mapLevel2;provinceC5001;trend(C1,C5001);/ 14.  Halfon P, Eggli Y, Van Melle G, Chevalier J, Wasserfallen JB, Burnand B. Measuring potentially avoidable hospital readmissions. J Clin Epidemiol. 2002;55(6):573-587. doi:10.1016/S0895-4356(01)00521-2 56  15.  Retrum JH, Boggs J, Hersh A, et al. Patient-identified factors related to heart failure readmissions. Circ Cardiovasc Qual Outcomes. 2013;6(2):171-177. doi:10.1161/CIRCOUTCOMES.112.967356 16.  Hu J, Gonsahn MD, Nerenz DR. Socioeconomic status and readmissions: Evidence from an urban teaching hospital. Health Aff. 2014;34(5):778-785. doi:10.1377/hlthaff.2013.0816 17.  Felix HC, Seaberg B, Bursac Z, Thostenson J, Stewart MK. Why Do Patients Keep Coming Back? Results of a Readmitted Patient Survey. Soc Work Health Care. 2015;54(1):1-15. doi:10.1080/00981389.2014.966881 18.  Brüngger B, Blozik E. Hospital readmission risk prediction based on claims data available at admission: A pilot study in Switzerland. BMJ Open. 2019;9(6):1-10. doi:10.1136/bmjopen-2018-028409 19.  Swanson JO, Vogt V, Sundmacher L, Hagen TP, Moger TA. Continuity of care and its effect on readmissions for COPD patients: A comparative study of Norway and Germany. Health Policy (New York). 2018;122(7):737-745. doi:10.1016/j.healthpol.2018.05.013 20.  Naylor M, Brooten D, Jones R, Lavizzo-Mourey R, Mezey M, Pauly M. Comprehensive discharge planning for the hospitalized elderly: A randomized clinical trial. Ann Intern Med. 1994;120(12):999-1006. doi:10.7326/0003-4819-120-12-199406150-00005 21.  Naylor MD. Comprehensive discharge planning for hospitalized elderly: A pilot study. Nurs Res. 1990;39(3):156-161. doi:10.1097/00004045-199009000-00011 22.  Brooten D, Brown LP, Munro BH, et al. Early Discharge and Specialist Transitional Care. Image J Nurs Scholarsh. 1988;20(2):64-68. doi:10.1111/j.1547-5069.1988.tb00032.x 23.  Kennedy L, Neidlinger S, Scroggins K. Effective Comprehensive Discharge Planning for 57  Hospitalized Elderly. Gerontologist. 1987;27(5):577-580. doi:10.1093/geront/27.5.577 24.  Berwick DM, Nolan TW, Whittington J. The triple aim: Care, health, and cost. Health Aff. 2008;27(3):759-769. doi:10.1377/hlthaff.27.3.759 25.  Hospital Readmissions Reduction Program (HRRP) | CMS. Accessed April 29, 2020. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Readmissions-Reduction-Program 26.  Banerjee S, Paasche-Orlow MK, McCormick D, Lin M-Y, Hanchate AD. Association between Medicare’s Hospital Readmission Reduction Program and readmission rates across hospitals by medicare bed share. BMC Health Serv Res. 2021;21(1):1-10. doi:10.1186/s12913-021-06253-2 27.  Ayabakan S, Bardhan I, Zheng Z (Eric). Triple Aim and the Hospital Readmission Reduction Program. Heal Serv Res Manag Epidemiol. 2021;8:233339282199370. doi:10.1177/2333392821993704 28.  Hellsten E, Liu G, Yue E, Gao G, Sutherland JM. Improving hospital quality through payment reforms: A policy impact analysis in British Columbia. Healthc Manag forum. 2016;29(1):33-38. doi:10.1177/0840470415614054 29.  Virtual Care Task Force. Virtual Care Recommendations for Scaling Up Virtual Medical Services.; 2020. https://www.cma.ca/sites/default/files/pdf/virtual-care/ReportoftheVirtualCareTaskForce.pdf 30.  Staples JA, Liu G, Brubacher JR, Karimuddin A, Sutherland JM. Physician Financial Incentives to Reduce Unplanned Hospital Readmissions: an Interrupted Time Series Analysis. Published online 2021. doi:10.1007/s11606-021-06803-8 31.  Lapointe-Shaw L, Mamdani M, Luo J, et al. Effectiveness of a financial incentive to 58  physicians for timely follow-up after hospital discharge: A population-based time series analysis. CMAJ. 2017;189(39):E1224-E1229. doi:10.1503/cmaj.170092 32.  Basnet S, Zhang M, Lesser M, et al. Thirty-day hospital readmission rate amongst older adults correlates with an increased number of medications, but not with Beers medications. Geriatr Gerontol Int. 2018;18(10):1513-1518. doi:10.1111/ggi.13518 33.  Garrison GM, Mansukhani MP, Bohn B. Predictors of thirty-day readmission among hospitalized family medicine patients. J Am Board Fam Med. 2013;26(1):71-77. doi:10.3122/jabfm.2013.01.120107 34.  Glans M, Kragh Ekstam A, Jakobsson U, Bondesson Å, Midlöv P. Risk factors for hospital readmission in older adults within 30 days of discharge – a comparative retrospective study. BMC Geriatr. 2020;20(1):1-12. doi:10.1186/s12877-020-01867-3 35.  Pedersen MK, Meyer G, Uhrenfeldt L. Risk factors for acute care hospital readmission in older persons in Western countries. JBI Database Syst Rev Implement Reports. 2017;15(2):454-485. doi:10.11124/JBISRIR-2016-003267 36.  Gershon AS, Thiruchelvam D, Aaron S, et al. Socioeconomic status (SES) and 30-day hospital readmissions for chronic obstructive pulmonary (COPD) disease: A population-based cohort study. PLoS One. 2019;14(5):1-17. doi:10.1371/journal.pone.0216741 37.  Kind AJH, Jencks S, Brock J, et al. Neighborhood socioeconomic disadvantage and 30-day rehospitalization: A retrospective cohort study. Ann Intern Med. 2014;161(11):765-774. doi:10.7326/M13-2946 38.  Brown AF, Liang LJ, Vassar SD, et al. Neighborhood socioeconomic disadvantage and mortality after stroke. Neurology. 2013;80(6):520-527. doi:10.1212/WNL.0b013e31828154ae 59  39.  Nagasako EM, Reidhead M, Waterman B, Claiborne Dunagan W. Adding socioeconomic data to hospital readmissions calculations may produce more useful results. Health Aff. 2014;34(5):786-791. doi:10.1377/hlthaff.2013.1148 40.  Murphy BM, Elliott PC, le Grande MR, et al. Living alone predicts 30-day hospital readmission after coronary artery bypass graft surgery. Eur J Prev Cardiol. 2008;15(2):210-215. doi:10.1097/HJR.0b013e3282f2dc4e 41.  Jasti H, Mortensen EM, Obrosky DS, Kapoor WN, Fine MJ. Causes and risk factors for rehospitalization of patients hospitalized with community-acquired pneumonia. Clin Infect Dis. 2008;46(4):550-556. doi:10.1086/526526 42.  Donze J, Aujesky D, Williams D, Schnipper JL. Potentially avoidable 30-day hospital readmissions in medical patients: Derivation and validation of a prediction model. JAMA Intern Med. 2013;173(8):632-638. doi:10.1001/jamainternmed.2013.3023 43.  van Walraven C, Wong J, Forster AJ. LACE+ index: Extension of a validated index to predict early death or urgent readmission after hospital discharge using administrative data. Open Med. 2012;6(2):1-11. Accessed May 31, 2021. /pmc/articles/PMC3659212/ 44.  Staples JA, Wiksyk B, Liu G, Desai S, van Walraven C, Sutherland JM. External validation of the modified LACE+, LACE+, and LACE scores to predict readmission or death after hospital discharge. J Eval Clin Pract. Published online 2021. doi:10.1111/jep.13579 45.  Maroko AR, Doan TM, Arno PS, Hubel M, Yi S, Viola D. Integrating Social Determinants of Health With Treatment and Prevention: A New Tool to Assess Local Area Deprivation. Prev Chronic Dis. 2016;13(9):160221. doi:10.5888/pcd13.160221 46.  Hu J, Kind AJH, Nerenz D. Area Deprivation Index Predicts Readmission Risk at an 60  Urban Teaching Hospital. Am J Med Qual. 2018;33(5):493-501. doi:10.1177/1062860617753063 47.  Chamberlain AM, Finney Rutten LJ, Wilson PM, et al. Neighborhood socioeconomic disadvantage is associated with multimorbidity in a geographically-defined community. BMC Public Health. 2020;20(1):13. doi:10.1186/s12889-019-8123-0 48.  Canadian Index of Multiple Deprivation: Dataset. Accessed May 3, 2021. https://www150.statcan.gc.ca/n1/pub/45-20-0001/452000012019002-eng.htm 49.  Division SM, Centre MI, Division AS. The Canadian Index of Multiple Deprivation. 2019;(Cimd). 50.  Schwartz S. The fallacy of the ecological fallacy: The potential misuse of a concept and the consequences. Am J Public Health. 1994;84(5):819-824. doi:10.2105/AJPH.84.5.819 51.  Shafran-Nathan R, Levy I, Levin N, Broday DM. Ecological bias in environmental health studies: the problem of aggregation of multiple data sources. doi:10.1007/s11869-016-0436-x 52.  Hansen LO, Young RS, Hinami K, Leung A WM. Interventions to Reduce 30-Day Rehospitalization : A Systematic Review. Ann Intern Med. 2011;155(8):520-528. 53.  Naylor MD, Sochalski JA. Scaling up: bringing the transitional care model into the mainstream. Issue Brief (Commonw Fund). 2010;103:1-12. Accessed February 1, 2021. https://pubmed.ncbi.nlm.nih.gov/21053533/ 54.  Naylor MD. A Decade of Transitional Care Research with Vulnerable Elders. J Cardiovasc Nurs. 2000;14(3):1-14. 55.  Fønss Rasmussen L, Grode LB, Lange J, Barat I, Gregersen M. Impact of transitional care interventions on hospital readmissions in older medical patients: A systematic review. 61  BMJ Open. 2021;11(1). doi:10.1136/bmjopen-2020-040057 56.  Herrin J, St. Andre J, Kenward K, Joshi MS, Audet AMJ, Hines SC. Community factors and hospital readmission rates. Health Serv Res. 2015;50(1):20-39. doi:10.1111/1475-6773.12177 57.  Shaw J, Jamieson T, Agarwal P, Griffin B, Wong I, Bhatia RS. Virtual care policy recommendations for patient-centred primary care: findings of a consensus policy dialogue using a nominal group technique. J Telemed Telecare. 2018;24(9):608-615. doi:10.1177/1357633X17730444 58.  Federation of Medical Regulatory Authorities of Canada. FMRAC Framework on telemedicine. 2020. Accessed April 16, 2021. http://fmrac.ca/fmrac-framework-on-telemedicine/ 59.  Connecting Patients for Better Health: 2018 | Canada Health Infoway. Accessed June 25, 2021. https://www.infoway-inforoute.ca/en/component/edocman/resources/reports/benefits-evaluation/3564-connecting-patients-for-better-health-2018 60.  Dorn SD. Backslide or forward progress? Virtual care at U.S. healthcare systems beyond the COVID-19 pandemic. npj Digit Med. 2021;4(1):1-3. doi:10.1038/s41746-020-00379-z 61.  WHO. mHealth: New horizons for health through mobile technologies. Observatory. 2011;3:102. http://www.who.int/goe/publications/goe_mhealth_web.pdf%5Cnhttp://www.who.int/goe/publications/ehealth_series_vol3/en/index.html 62.  Nurgalieva L, O’Callaghan D, Doherty G. Security and Privacy of mHealth Applications: A Scoping Review. IEEE Access. 2020;8:104247-104268. 62  doi:10.1109/ACCESS.2020.2999934 63.  Canadian Radio-Television and Telecommunications Commission. Broadband Fund: Closing the digital divide in Canada. Published online 2020. Accessed April 16, 2021. https://crtc.gc.ca/eng/internet/internet.htm 64.  Arora S, Peters AL, Burner E, Lam CN, Menchine M. Trial to examine text message-based mhealth in emergency department patients with diabetes (TExT-MED): A randomized controlled trial. Ann Emerg Med. 2014;63(6):745-754.e6. doi:10.1016/j.annemergmed.2013.10.012 65.  Jennings L, Lee N, Shore D, et al. U.S. Minority Homeless Youth’s Access to and Use of Mobile Phones: Implications for mHealth Intervention Design. J Health Commun. 2016;21(7):725-733. doi:10.1080/10810730.2015.1103331 66.  Larbi D, Randine P, Årsand E, Antypas K, Bradway M, Gabarron E. Methods and evaluation criteria for apps and digital interventions for diabetes self-management: Systematic review. J Med Internet Res. 2020;22(7):1-13. doi:10.2196/18480 67.  Wang Y, Xue H, Huang Y, Huang L, Zhang D. A Systematic Review of Application and Effectiveness of mHealth Interventions for Obesity. Adv Nutr. 2017;8(3):449-462. doi:10.3945/an.116.014100.449 68.  Stevenson JK, Campbell ZC, Webster AC, et al. Ehealth interventions for people with chronic kidney disease. Cochrane Database Syst Rev. 2019;2019(8). doi:10.1002/14651858.CD012379.pub2 69.  Choi WS, Choi JH, Oh J, Shin IS, Yang JS. Effects of Remote Monitoring of Blood Pressure in Management of Urban Hypertensive Patients: A Systematic Review and Meta-Analysis. Telemed e-Health. 2020;26(6):744-759. doi:10.1089/tmj.2019.0028 63  70.  Lester RT, Ritvo P, Mills EJ, et al. Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): A randomised trial. Lancet. 2010;376(9755):1838-1845. doi:10.1016/S0140-6736(10)61997-6 71.  Patel AR, Campbell JR, Sadatsafavi M, et al. Burden of non-adherence to latent tuberculosis infection drug therapy and the potential cost-effectiveness of adherence interventions in Canada: A simulation study. BMJ Open. 2017;7(9). doi:10.1136/bmjopen-2016-015108 72.  A Randomized Controlled Trial of a Text Messaging Based Intervention Versus a Written Action Plan in Asthma Management: Results from a Feasibility Study | A94. IMPROVING ASTHMA MANAGEMENT: RESEARCH AT THE FOREFRONT. Accessed April 8, 2021. https://www.atsjournals.org/doi/abs/10.1164/ajrccm-conference.2017.195.1_MeetingAbstracts.A2637 73.  Haida Gwaii project allows patients to reach health-care professionals via text | CBC News. Accessed April 8, 2021. https://www.cbc.ca/news/canada/british-columbia/haida-gwaii-project-allows-patients-to-reach-health-care-professionals-via-text-1.4074368 74.  Jongbloed K, Pearce ME, Thomas V, et al. The cedar project - Mobile phone use and acceptability of mobile health among young indigenous people who have used drugs in British Columbia, Canada: Mixed methods exploratory study. JMIR mHealth uHealth. 2020;8(7). doi:10.2196/16783 75.  Chiang N, Guo M, Amico KR, Atkins L, Lester RT. Interactive two-way mhealth interventions for improving medication adherence: An evaluation using the behaviour change wheel framework. JMIR mHealth uHealth. 2018;6(4). doi:10.2196/mhealth.9187 76.  Jones PH, Shakdher S, Singh P. Synthesis maps: Visual knowledge translation for the 64  canIMPACT clinical system and patient cancer journeys. Curr Oncol. 2017;24(2):129-134. doi:10.3747/co.24.3452 77.  Edgcomb JB, Zima B. Machine learning, natural language processing, and the electronic health record: Innovations in mental health services research. Psychiatr Serv. 2019;70(4):346-349. doi:10.1176/appi.ps.201800401 78.  Yim WW, Yetisgen M, Harris WP, Sharon WK. Natural Language Processing in Oncology Review. JAMA Oncol. 2016;2(6):797-804. doi:10.1001/jamaoncol.2016.0213 79.  Liu Y, Chen PHC, Krause J, Peng L. How to Read Articles That Use Machine Learning: Users’ Guides to the Medical Literature. JAMA - J Am Med Assoc. 2019;322(18):1806-1816. doi:10.1001/jama.2019.16489 80.  Venkatramanan S, Sadilek A, Fadikar A, et al. Forecasting influenza activity using machine-learned mobility map. doi:10.1038/s41467-021-21018-5 81.  Alessa A, Faezipour M. Preliminary Flu Outbreak Prediction Using Twitter Posts Classification and Linear Regression With Historical Centers for Disease Control and Prevention Reports: Prediction Framework Study. doi:10.2196/12383 82.  College of Family Physicians of Canada. A new vision for Canada: Family Practice—The Patient’s Medical Home. Published online 2019. 83.  Radio-television C, Commission T. Communications Monitoring Report. Accessed April 17, 2021. https://applications.crtc.gc.ca/contact/eng/library 84.  Campbell AR, Kinvig K, Côté HCF, et al. Health care provider utilization and cost of an mHealth intervention in vulnerable people living with HIV in Vancouver, Canada: Prospective study. JMIR mHealth uHealth. 2018;6(7):1-13. doi:10.2196/mhealth.9493 85.  Peiris D, Sun L, Patel A, et al. Systematic medical assessment, referral and treatment for 65  diabetes care in China using lay family health promoters: protocol for the SMARTDiabetes cluster randomised controlled trial. Published online 2016. doi:10.1186/s13012-016-0481-8 86.  Alluhaidan A, Lee E, Alnosayan N, et al. Designing patient-centered mHealth technology intervention to reduce hospital readmission for heart-failure patients. In: Proceedings of the Annual Hawaii International Conference on System Sciences. Vol 2015-March. IEEE Computer Society; 2015:2886-2895. doi:10.1109/HICSS.2015.349 87.  Schwamm LH. Telehealth: Seven strategies to successfully implement disruptive technology and transform health care. Health Aff. 2014;33(2):200-206. doi:10.1377/hlthaff.2013.1021 88.  Mueller EL, Cochrane AR, Bennett WE, Carroll AE. A survey of mobile technology usage and desires by caregivers of children with cancer. Pediatr Blood Cancer. 2018;65(11):1-7. doi:10.1002/pbc.27359 89.  Sánchez-Cabrero R, Costa-Román Ó, Pericacho-Gómez FJ, Novillo-López MÁ, Arigita-García A, Barrientos-Fernández A. Early virtual reality adopters in Spain: sociodemographic profile and interest in the use of virtual reality as a learning tool. Heliyon. 2019;5(3). doi:10.1016/j.heliyon.2019.e01338 90.  BC Community Health Data. Accessed April 29, 2021. http://communityhealth.phsa.ca/HealthProfiles/HealthReport/Vancouver - Aggregate?archiveYear=2019 91.  False Discovery Rate | Columbia Public Health. Accessed June 24, 2021. https://www.publichealth.columbia.edu/research/population-health-methods/false-discovery-rate 66  92.  Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J R Stat Soc Ser B. 1995;57(1):289-300. doi:10.1111/j.2517-6161.1995.tb02031.x 93.  Curtis JR, Nielsen EL, Treece PD, et al. Effect of a Quality-Improvement Intervention on End-of-Life Care in the Intensive Care Unit A Randomized Trial. Am J Respir Crit Care Med. 2011;183:348-355. doi:10.1164/rccm.201006-1004OC 94.  Cavalcanti AB, Bozza FA, Machado FR, et al. Effect of a quality improvement intervention with daily round checklists, goal setting, and clinician prompting on mortality of critically ill patients: A randomized clinical trial. JAMA - J Am Med Assoc. 2016;315(14):1480-1490. doi:10.1001/jama.2016.3463 95.  Nuckols TK, Keeler E, Morton S, et al. Economic evaluation of quality improvement interventions designed to prevent hospital readmission: A systematic review and meta-analysis. JAMA Intern Med. 2017;177(7):975-985. doi:10.1001/jamainternmed.2017.1136 96.  Seeman S. Home Is Best - Vancouver Coastal Health. Accessed April 8, 2021. www.vch.ca 97.  Home support - Vancouver Coastal Health. Accessed April 8, 2021. http://www.vch.ca/your-care/home-community-care/care-options/home-support 98.  Making the transition home - Vancouver Coastal Health. Accessed April 8, 2021. http://www.vch.ca/about-us/news/making-the-transition-home 99.  Vancouver Coastal Health, Providence Health Care. Our Health Care Report Card. Published online 2020:1. 100.  Hoffman GJ, Yakusheva O. Association Between Financial Incentives in Medicare’s Hospital Readmissions Reduction Program and Hospital Readmission Performance. 67  JAMA Netw open. 2020;3(4):e202044. doi:10.1001/jamanetworkopen.2020.2044 101.  Foley T, Fairmichael F, Org WL. The Potential of Learning Healthcare Systems Executive Overview.; 2015. Accessed May 31, 2021. www.learninghealthcareproject.org 102.  Obermeyer Z, Emanuel EJ. Predicting the Future — Big Data, Machine Learning, and Clinical Medicine. N Engl J Med. 2016;375(13):1216-1219. doi:10.1056/nejmp1606181 103.  Alberta Health Services. Toward Optimized Practice. Benefits of a Patient’s Medical Home–A Literature Summary of 115 Articles.; 2017. https://actt.albertadoctors.org/file/top--evidence-summary--benefits-of-pmh.pdf 104.  Zheng S, Hanchate A, Shwartz M. One-year costs of medical admissions with and without a 30-day readmission and enhanced risk adjustment. BMC Health Serv Res. 2019;19(1):1-10. doi:10.1186/s12913-019-3983-7 105.  Anderson GF, Steinberg EP. Hospital Readmissions in the Medicare Population. N Engl J Med. 1984;311(21):1349-1353. doi:10.1056/NEJM198411223112105 106.  Canadian Medical Association. What Canadians Think About Virtual Health Care.; 2020. https://www.cma.ca/sites/default/files/pdf/virtual-care/cma-virtual-care-public-poll-june-2020-e.pdf  68  Appendices Appendix A  Readmission Risk Assessment Score (RRAS)      69  Appendix B  My Discharge Plan   70      71  Appendix C  Contingency Tables Table C.1 Contingency Table of Patients’ Internet Access by SMS Use SMS use Internet Access  Total No Yes No 9 75% 3 8.3% 12 25% Yes 3 25% 33 91.7% 36 75% Total 12 100.0% 36 100.0% 48 p-value = 2.32E-5  Table C.2 Contingency Table of Patients’ Type of Phone by Internet Access Internet Access Type of Phone Total Other Smart phone No 9 81.8% 3 8.3% 12 25.53 Yes 2 18.2% 33 91.7% 35 72.92% Total 11 100.0% 36 100.0% 47 p-value = 7.65E-6  Table C.3 Contingency Table of Patients’ Internet Access by Interest in Opportunity to Text Health Care Provider (HCP) Would Like Opportunity to Text HCP Internet Access Total No Yes No 6 50% 6 16.7% 12 25% Yes 6 50% 30 83.3% 36 75% Total 12 100.0% 36 100.0% 48 p-value = 0.049 72  Table C.4 Contingency Table of Patients’ Mobile Phone Access via Proxy by Interest in Opportunity to Text Health Care Provider (HCP) Would Like Opportunity to Text HCP  Access via Proxy Total No Yes No 8 80% 2 12.5% 10 38.46 Yes 2 20% 14 87.5% 16 61.54% Total 10 100.0% 16 100.0% 26 p-value = 0.001  Table C.5 Contingency Table of Patients’ Previous SMS Communication with HCP by Interest in Opportunity to Text Health Care Provider (HCP) Would Like Opportunity to Text HCP  Ever Texted HCP Total No Yes No 19 34.5% 1 5.9% 20 27.8% Yes 36 65.5% 16 94.1% 52 72.2% Total 55 100.0% 17 100.0% 72 p-value = 0.028  Table C.6 Contingency Table of Patients’ Type of Phone by Interest in Opportunity to Text Health Care Provider (HCP) Would Like Opportunity to Text HCP  Type of Phone Total Other Smart phone No 6 54.5% 6 16.2% 12 25% Yes 5 45.5% 31 83.8% 36 75% Total 12 100.0% 37 100.0% 48 p-value = 0.018 73  Table C.7 Contingency Table of Patients’ Type of Phone by SMS Use SMS Use Type of Phone Total Other Smart phone No 7 63.6% 4 10.8% 11 22.92% Yes 4 36.4% 33 89.2% 37 77.08% Total 11 100.0% 37 100.0% 48 p-value = 0.001                  74  Appendix D  Patient Mobile Phone Access Survey 1. Do you own a cell phone? o Yes (1)  o No (2)   2. Is there someone who can send and receive text messages on your behalf?  o Yes (1)  o No (2)   3. If so, what is your relationship to that person (e.g., family, friend…)? ___________________  4. Is it a shared phone or a personal phone? o Shared (1)  o Personal (2)  o Other (3)   5. Who is it shared with? e.g., Friend, Family, Spouse, other...Explain 'Other': ______________ 6. What type of phone do you use? o Basic Phone (text/call) (1)  o Features Phone (text/call/internet) (2)  o Smart Phone (IOS) (3)  o Smart Phone (Android) (4)  o Smart Phone (Other) (5)   75  7. Does it come with a plan where you can call/text? o Yes, text and call (1)  o Call only (2)  o Text only (3)  o No (4)   8. Do you have internet access on your phone? o Data + Wi-Fi (1)  o Wi-Fi only (2)  o None (3)   9. Do you text message on your phone? o Yes (1)  o No (2)   10. What communication method do you use most often? (Please Rank) ______ Voice (1) ______ Video (2) ______ Text (3)    76  11. Have you ever texted with your health care provider (e.g., doctor or nurse)? o Yes (1)  o No (2)   12. Would you like the opportunity to text your HCP? (Choose all that apply) ▢ Yes, for appointment reminders (one-way) (1)  ▢ Yes, to receive health information (one-way) (2)  ▢ Yes, for medication reminders (one-way) (3)  ▢ Yes, for medication monitoring (two-way, e.g., to discuss side effects, prescription refills)   ▢ Yes, to discuss healthcare concerns (two-way) (5)  ▢ Yes, for other reasons (please specify) (6)  ▢ No (7)      77  Appendix E  H2home Post-Survey Note: When we say “you” or “your” in this survey, we mean the patient. 1. Are you the patient or filling out this survey on behalf of the patient? • I am the patient • I am filling out this survey on behalf of the patient  2. Did you revisit a hospital Emergency Department since you were discharged after your enrollment into this study? • Yes o Number of times: _________ o Reason: ___________ o Location: VGH or other: ____________ • No  3. Did you get admitted to hospital since you were discharged after your enrollment into this study? • Yes o Number of times: _________ o Reason: ___________ o Location: VGH or other: ____________ • No  4. Did you feel that WelTel helped you navigate your care since being discharged after your enrollment into this study? • Yes • No  5. What did you text about with your healthcare provider? (Select all that apply) • Response to check-ins  • Appointment reminders  • Health concerns  • Medication dosage  • Side effects • Other: _______________________________________________________________  6. If you had a health-related problem or concern, what was your first mode of communication with your healthcare provider to discuss the issue? (check all that apply)  • Texting via WelTel • Calling healthcare provider directly • Visiting the hospital or clinic directly • Other: __________________________________________________________________  78  7. What do you value about texting through WelTel? Include as many values as you like and indicate which is most important to you. __________________________________________  8. What challenges did you experience when texting through WelTel, if any? List as many challenges as you like and indicate which is most important to you. ____________________  9. Can you access the internet on your phone? • Yes • No  10. The WelTel platform includes video-calling service. Did you video with a healthcare provider?  • Yes • No  11. Would you be interested in using the video service?  • Yes • No  In the future, how would you prefer to communicate with your healthcare provider about:  12. Appointment reminders? Check all that apply.   • Texting  • Calling  • Mail  13. Check-ins on how you are doing between appointments? Check all that apply.  • Texting  • Calling  • Video  14. Would you use this service again to communicate with your healthcare provider? • Yes • No • Explain: ________________________________________________________________  15. How many times did you access your discharge plan via the WelTel platform? • 0 • 1-2 • 2-3 • 4 or more  16. How did it impact your care (e.g., transition back to the community)? (You may give an example) _______________________________________________________________ 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0400040/manifest

Comment

Related Items