Open Collections

UBC Graduate Research

Defining frailty within the context of electronic medical records : eCGA mapping to SNOMED CT Brown, Sharde 2019

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
42591-Brown_S_NURS595_eCGA_data_mapping_2019.pdf [ 892.6kB ]
Metadata
JSON: 42591-1.0377757.json
JSON-LD: 42591-1.0377757-ld.json
RDF/XML (Pretty): 42591-1.0377757-rdf.xml
RDF/JSON: 42591-1.0377757-rdf.json
Turtle: 42591-1.0377757-turtle.txt
N-Triples: 42591-1.0377757-rdf-ntriples.txt
Original Record: 42591-1.0377757-source.json
Full Text
42591-1.0377757-fulltext.txt
Citation
42591-1.0377757.ris

Full Text

DEFINING FRAILTY WITHIN THE CONTEXT OF ELECTRONIC MEDICAL RECORDS: eCGA DATA MAPPING TO SNOMED CT by  Sharde Marie Brown  BSN, Kwantlen Polytechnic University, 2009  A SCHOLARY PRACTICE ADVANCEMENT RESEARCH PAPER SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE  in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Nursing)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   March, 2019  © Sharde Marie Brown, 2019 ii  Abstract Background. Frailty is commonly considered an emerging geriatric syndrome that involves a decrease in the physiological ability to respond to stressors, which can lead to increased hospitalization, worsening health, and morbidity. The goals of this work are to: (1) identify frailty within primary health care as early as possible in order to prevent worsening health issues, and (2) explore information collected within electronic medical records that can be better utilized for assessing frailty. Specifically, this work: 1) conducted a literature review to identify what defining characteristics of frailty are captured in the CGA tool; and 2) matched concepts within the Comprehensive Geriatric Assessment (CGA) tool to SNOMED CT, an international standardized terminology.  Methods. This research was conducted using a nonexperimental descriptive study design. Manual mapping and expert consensus mapping were used to match concepts from a validated frailty assessment tool the electronic CGA (eCGA) to SNOMED CT; standardized terminologies allow for the sharing and coding of clinical information across different platforms. A literature search was completed to define the relevant characteristics of frailty, and explore the most commonly used assessment tools. These findings were compared with the eCGA to assess content coverage. A visual map supported mapping activities. Results and Conclusion. The literature review showed that the eCGA contained all defining characteristics of frailty. One hundred and thirty-three unique assessment questions were manually mapped to SNOMED CT. Of these, 72% (96/133) were direct matches, 17% (22/133) were one-to-many matches to clinical terms within SNOMED CT, and the remaining 11% (15/133) were non-matches. Two rounds of expert clinician consensus mapping occurred. Inter-rater reliability between the two clinicians was 0.75 (Kappa). The outcome of this study is a list iii  of mapped eCGA elements to SNOMED CT. This list informs a new data table added to a pan-national database for chronic disease monitoring. iv  Table of Contents  Abstract .......................................................................................................................................... ii Table of Contents ......................................................................................................................... iv List of Tables .............................................................................................................................. viii List of Figures ............................................................................................................................... ix List of Abbreviations .....................................................................................................................x Acknowledgements ...................................................................................................................... xi Dedication .................................................................................................................................... xii Chapter 1: Introduction ..............................................................................................................13 1.1 Background ................................................................................................................... 14 1.1.1 Primary Health Care ................................................................................................. 14 1.2 Research Questions ....................................................................................................... 16 1.3 Frailty and Aging .......................................................................................................... 17 1.3.1 Defining Frailty ......................................................................................................... 17 1.3.2 Assessing Frailty ....................................................................................................... 19 1.3.2.1 Assessment of Frailty in Primary Health Care ................................................. 20 1.3.3 Use of EMR for Frailty Research Advancement ...................................................... 20 1.3.3.1 Canadian Primary Care Sentinel Surveillance Network (CPCSSN) ................ 21 1.4 Conclusion .................................................................................................................... 22 Chapter 2: Literature Review .....................................................................................................23 2.1 Introduction ................................................................................................................... 23 2.2 Literature Identification ................................................................................................ 23 v  2.3 Validated Frailty Assessment Measures ....................................................................... 24 2.3.1 Frailty Index .............................................................................................................. 26 2.3.2 Frailty Phenotype ...................................................................................................... 27 2.3.3 Clinical Frailty Scale................................................................................................. 27 2.4 Frailty Assessment in Primary Care ............................................................................. 28 2.4.1 CARES Model eCGA ............................................................................................... 29 2.5 Electronic Medical Record Use in Canadian Primary Health Care .............................. 30 2.6 Standardized Terminologies ......................................................................................... 32 2.6.1 Standardized Clinical Terminologies ........................................................................ 34 2.6.1.1 Systematized Nomenclature of Medicine Clinical Term .................................. 35 2.6.1.2 Other Standardized Terminologies ................................................................... 35 2.6.2 Clinical Terminology Mapping................................................................................. 36 2.6.3 Clinical Impact of Standardized Terminology Usage ............................................... 37 2.7 Summary ....................................................................................................................... 38 Chapter 3: Methods .....................................................................................................................39 3.1 Introduction ................................................................................................................... 39 3.2 Research Questions ....................................................................................................... 39 3.3 Study Design ................................................................................................................. 39 3.4 Sampling Plan ............................................................................................................... 40 3.5 Procedure and Data Collection ..................................................................................... 41 3.6 Phase 1: The Defining Characteristics of Frailty .......................................................... 41 3.7 Phase 2: eCGA Data Sources and Concepts Map ......................................................... 42 3.8 Phase 3: Manual Mapping to SNOMED CT ................................................................ 43 vi  3.9 Phase 4: Clinician Consensus Mapping ........................................................................ 46 3.10 Data Analysis ................................................................................................................ 47 Chapter 4: Results........................................................................................................................48 4.1 Phase 1: The Defining Characteristics of Frailty .......................................................... 48 4.2 Phase 2: eCGA Map of Data Sources and Concepts .................................................... 51 4.3 Phase 3: Manual Mapping to SNOMED CT ................................................................ 55 4.4 Phase 4: Phase 4: Clinician Consensus Mapping ......................................................... 55 4.5 Summary ....................................................................................................................... 56 Chapter 5: Discussion ..................................................................................................................57 5.1 Introduction ................................................................................................................... 57 5.2 Defining Frailty ............................................................................................................. 57 5.3 Implications of Defining Frailty Through the eCGA ................................................... 58 5.4 Manual Standardized Terminology Mapping ............................................................... 59 5.4.1 Lessons Learned from Manual Mapping .................................................................. 60 5.5 Expert Consensus Mapping .......................................................................................... 62 5.6 Implications................................................................................................................... 63 5.6.1 Clinical Workforce Education .................................................................................. 64 5.6.2 Data Collection Quality ............................................................................................ 65 5.6.3 Mapping Research .................................................................................................... 66 5.7 Limitations .................................................................................................................... 67 5.8 Recommendations ......................................................................................................... 68 5.8.1 CPCSSN Database and Recommendations .............................................................. 69 5.8.2 Canadian Primary Care Sentinel Surveillance Network Data .................................. 70 vii  5.9 The Role of Nursing in PHC and Frailty Research ....................................................... 71 5.10 Summary ....................................................................................................................... 73 References .....................................................................................................................................74 Appendices ....................................................................................................................................88 Appendix A CARES CGA Questionnaire ................................................................................ 88 Appendix B Manual Mapping .................................................................................................. 89 Appendix C Clinician Consensus #1 ........................................................................................ 94 Appendix D Clinician Consensus #2 ........................................................................................ 96    viii  List of Tables Table 3.1 Clinical term mapping matching criteria ...................................................................... 45 Table 3.2 Clinician mapping term matching criteria .................................................................... 47 Table 4.1 Defining Characteristics of Frailty ............................................................................... 48 Table 4.2 Characteristics of Frailty in Assessment Tools............................................................. 49 Table 4.3 Legend-Overall Map of Data Sources and Concepts .................................................... 51 Table 4.4 Manual Mapping of eCGA to SNOMED CT ............................................................... 55 Table 4.5 Clinician Consensus Mapping of eCGA to SNOMED CT- Iteration #1 ...................... 56 Table 4.6 Clinician Consensus Mapping of eCGA to SNOMED CT- Iteration #2 ...................... 56  ix  List of Figures Figure 3.1 Overview of Study Method ......................................................................................... 40 Figure 4.1 Overall Map of Data Sources and Concepts ................................................................ 52 Figure 4.2 Data Sources ................................................................................................................ 53 Figure 4.3 Exam Data ................................................................................................................... 54                   x  List of Abbreviations CFS  ............................ Clinical Frailty Scale CGA  ........................... Comprehensive Geriatric Assessment CIHI ............................ Canadian Institute for Health Information CMA  .......................... Canadian Medical Association CPCSSN  ..................... Canadian Primary Care Sentinel Surveillance Network  eCGA  ......................... Electronic Comprehensive Geriatric Assessment EMR  ........................... Electronic Medical Record FI  ................................ Frailty Index ICD  ............................. International Classification of Diseases LOINC  ....................... Logical Observation Identifiers Names and Codes PCP  ............................ Primary Care Practitioner PHC  ............................ Primary Health Care SNOMED CT ............. Systematized Nomenclature of Medicine Clinical Terms xi  Acknowledgements I would like to acknowledge my supervisor Dr. Sabrina Wong and committee member Dr. Leanne Currie for their ongoing support, guidance, and encouragement. I would like to thank Manpreet Gill Thandi who assisted me with the clinician validation portion of this paper, and answered numerous questions regarding the application of the eCGA form. I would also like to offer a special thank you to the data managers of the CPCSSN database, who also assisted me in this endeavor.   Without the undying support from my family, completing this research project would not have been possible. I would like to thank my husband and my mother for allowing me the space to achieve my goals. xii  Dedication  I would like to dedicate this paper to my two daughters. I hope that I can inspire you to pursue your educational goals, no matter what stage of life you are in. I hope that I have showed you that you can do anything you set your mind to. I love you both more than you will know.              13  Chapter 1: Introduction In the last decade frailty has become an increasingly discussed topic within the health care research community (Brown & Covinsky, 2018). The research to date has focused on attempting to provide a concise operational definition of frailty that can be applied in a number of clinical settings. This includes a definition of frailty, the biological components and markers that contribute to this condition, and the consequences that arise when frailty has been ascertained in an individual (Harmand et al., 2017). Frailty is defined as a “state of increased vulnerability, with reduced physiological reserve and loss of function across multiple body systems” (Canadian Frailty Network, 2017a). As multiple physiological systems decline, the extent to which frailty manifests depends on the body’s cellular mechanisms to repair and maintain homeostasis (Dent, Kowal, & Hoogendijk, 2016). People living with frailty have a reduced ability to cope with nominal stressors which can lead to rapid changes within the health of that individual  (Canadian Frailty Network, 2017a) .  While frailty does not exclusively affect people of older age, people older than 65 years currently account for the biggest proportion of the population affected (Fried et al., 2001). There are many adverse outcomes for people living with frailty that range from increased hospital stay, mortality rates and falls, to decreased mobility and overall worsening health status (Fried et al., 2001; Kenneth Rockwood & Mitnitski, 2011). The resulting factors of increased frailty can place a burden on individuals, families and caretakers, and the health care system at large. People who are frail are at greater risk for chronic illnesses, and disabilities which impact daily living activities ( Van Velsen et al., 2015; Fugate Woods et al., 2005). The Canadian health care system spends upwards of $90 billion dollars on treatment of chronic illnesses (Canadian Institute for Health Information, 2016) and early identification and intervention for frailty may help reduce 14  these costs. Multifactorial and physical therapy related intervention results provide some evidence that they can reduce the severity of frailty in individuals (Romera-Liebana et al., 2018; Theou et al., 2017). The goal for early intervention and treatment of frailty should be to improve overall quality of life, compress the timeframe of morbidity, and improve the end stages of life.  1.1 Background In Canada, people over the age of 65 will account for approximately 25% of the population by the year 2030 (Bohnert, Chagnon, & Dion, 2015). At present, there are approximately 1.2 million Canadians living with frailty, with that number expected to increase to more than 2 million by 2035 (Canadian Frailty Network, 2017). With the expected increase in aged people, it is important to understand the impact this may have on both the healthcare system and family members who help care for their aging relatives, and identify prevention strategies if possible. As healthcare providers, it is important to understand how we can better identify and assist this growing population. If there are better tools to help identify frailty in a variety of settings, earlier interventions to mitigate some of the adverse outcomes associated with frailty may be implemented. This is especially true in the primary health care setting, because this setting is often the first and main point of contact for patients and families trying to navigate the Canadian healthcare system.   1.1.1 Primary Health Care Primary health care (PHC) is defined as “Care which provides integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context 15  of family and community” (Vanselow, 1995). Based on 2014 survey, it is estimated that over 80% of those aged 65 years and older in Canada have access to PHC (Canadian Institute for Health Information, 2016). In Canada, PHC services are typically delivered through a “most responsible person” which is usually a nurse practitioner or family physician. These services include, but are not limited to, routine health care as well as health promotion and disease prevention (Canadian Institute for Health Information, 2017). Because many people aged 65 years and older currently use PHC as the first point of access to healthcare, it is imperative that clinicians are provided with the tools and knowledge needed to help promote best practices to the growing population.  The Canadian Medical Association identified the need for the development of a national seniors’ strategy that would address some of the current constraints while planning for the future of  health care services for seniors (Canadian Medical Association, 2016). One of these strategies would be providing “improved training, resource allocation and incentives to help primary care practitioners (PCP) develop robust, around-the-clock services for frail and elderly Canadians living in the community” (Canadian Medical Association, 2016). This is especially important as currently there are only 294 Geriatricians, and 192 Geriatric Psychiatrists serving all of Canada (Canadian Medical Association, 2017). By providing PCPs with more training in frailty assessment and treatment, the limited number of Geriatricians could focus their practice on more complex patient populations, such as people suffering from multi-morbidities and degenerative diseases.  Given the need to reliably assess for frailty in PHC, and subsequently develop effective interventions, we need to better understand how to effectively utilize the existing tools available in the current practice environment. In part, this involves understanding key defining 16  characteristics of frailty and how the PHC clinician can use this information for further assessment and care planning.    1.2 Research Questions This project addresses two research questions: Research Question 1: What are the defining characteristics of frailty? Are these characteristics captured in the electronic Comprehensive Geriatric Assessment (eCGA) form? Research Question 2: What data elements from the eCGA can be mapped to existing SNOMED CT standardized terminology and what is the rate of equivalence?  The goals of this work are to: (1) identify frailty within PHC since it an ideal setting to identify and prevent worsening health issues, and (2) explore information collected within electronic medical records (EMRs) that can be better utilized for assessing frailty. Recent initiatives to increase the use of EMRs within PHC, standardize data entry, and provide networks in which to extract those data can be a source of clinical data for researchers and inform the referral process from PHC to specialized geriatricians.  This project identifies defining characteristics of frailty using a literature review and mapping process. It also identifies specific data elements through standard terminology mapping within EMRs. The ability of PCPs to accurately identify frailty within their patient populations could be used to promote better patient outcomes through health promotion, prevention, and early intervention. The identification, counselling, and preventative measures that can be taken on an individual basis will hopefully prevent future adverse outcomes associated with frailty. Finally, the accurate identification of frailty in PHC can improve the referral process to specialty care.  17  1.3 Frailty and Aging While frailty primarily affects individuals of older chronological age, there are differentiations between a state of frailty and the normal aging process (Fedarko, 2011). Aging is a process that inevitably occurs, including a progressive decline and deterioration of the physiological properties at the cellular, tissue, and organ level resulting in death. Frailty is defined as the increased vulnerability and decreased physiological reserves to be able to respond to stressors, which result in a loss of function across multiple body systems (Canadian Frailty Network, 2017). Both the normal aging process and frailty have similarities and include a loss of homeostasis. While aging generally happens systemically and progressively, frailty involves the faster failure of homeostatic stability in energy metabolism and neuromuscular abilities which lead to a variety of health-related deficits (Fedarko, 2011). For example, a person who is 80 years old, is physically and socially active, with no cognitive declines may not be frail, whereas a 70 year old person with mobility issues, frequent episodic and chronic health problems might be classified as frail. To understand frailty, it is important to decouple the idea that age and frailty directly correlate with one another.   1.3.1 Defining Frailty In the current available literature, there are multiple ways of defining frailty. One of the main objectives of current research regarding frailty involves attempts to provide one working operational definition of frailty that would satisfy health care clinicians as well as researchers. However, this has proven to be unsuccessful as the concept of frailty remains somewhat elusive. Frailty is not a diagnosable “disease state”, does not have specific biomedical markers and remains conceptually complex. Thus, the tools currently being used to assess frailty vary 18  depending on the context in which frailty is being discussed (Harmand et al., 2017); Some of the research suggests that frailty is a syndrome (Xue, 2011), while other researchers define it as being a chronic condition (Fried, Ferrucci, Darer, Williamson, & Anderson, 2004). Also, there are some competing thoughts as to whether or not frailty (as a condition) should include things like comorbidities and disabilities as predictors and inclusion factors in the assessment of an individual (Cesari, Calvani, & Marzetti, 2017). There have been studies showing that disability and comorbid conditions are not necessarily an inclusion factor in patients presenting as frail (Fried, et al., 2004). However there have been more recent studies that suggest comorbidity and disability overlap with frailty, especially as the severity of frailty increases (Theou, Rockwood, Mitnitski, & Rockwood, 2012). Much of the discourse related to defining frailty relates to the inclusion of specific diagnostic criteria within the definition that would make it universally accepted. Rodríguez-Mañas et al., (2013) used a Delphi method to gain consensus among a panel of experts. While their findings suggest there is agreement on the concept of frailty, a lack of consensus remains about which biophysical markers should be included within this definition and if other stressors such as psychosocial and socio-economic factors should also be assessed. There is also discourse regarding a method to assess frailty that would include the severity of these markers. Experts also disagreed about a clear chronological timeline for assessing for frailty within the patient population. This lack of consensus leaves room for ambiguity and fails to provide a gold standard definition with which to further research in this area. If there is not one agreed upon definition, it makes it difficult to compare or replicate research or to implement quality improvement studies. This is especially true for the PHC clinician who relies on current best evidence to guide their practice. The definition of frailty that is used in this project is based on the definition from the 19  Canadian Frailty Network, which is comprised of health care and research networks across Canada, and is funded by the Government of Canada’s Networks of Centres of Excellence program (Canadian Frailty Network, 2017). The goal of the CFN is to develop programs and products, through research and partnership that would improve the quality of economic outcomes for the Canadian health care system, as well as improve the quality of life for Canadians and families living with frailty (Canadian Frailty Network, 2017).  1.3.2 Assessing Frailty While there is a lack of a single operational definition of frailty, there are also numerous ways in which frailty can be assessed for any given population. This has presented a problem in being able to gain consensus among health care providers, and advance research into the effectiveness of frailty prevention and treatment. Many assessment tools have been developed within the last 15 years. Currently, there are over 40 tools designed to assess frailty, with more being developed as research continues to expand in this area. Many of the frailty assessment tools have not been studied beyond their initial research papers (Theou, Walston, & Rockwood, 2015). Each of these types of assessment and screening tools claims to be able to identify frail subjects both individually and as a cohort as well as retrospectively through data aggregation. The tools differ in their inclusion of variables and outcomes (Cesari, Calvani, & Marzetti, 2017a). The assessment tools also range from identifying frailty as being present or absent, to being able to provide a range of severity from pre-frail to severely frail (Cesari, Gambassi, Van Kan, & Vellas, 2014). Having this variety in frailty assessment tools, inclusion criteria, and diagnosis outcomes, makes it difficult to pinpoint a single tool to recommend for widespread implementation.  20  1.3.2.1 Assessment of Frailty in Primary Health Care Because of the lack of evidence, routine frailty assessment for all elderly patients in PHC is not currently recommended by the BC Ministry of Health (Ministry of Health, 2017). Instead, the Ministry recommends using a diligent case finding approach to identify patients with frailty, particularly among older adults who regularly or increasingly require health and social services (Ministry of Health, 2017). Currently, there is not one specific tool validated for use within the PHC practice environment. With the amount of frailty related research available, it has been difficult to synthesize and translate the findings from the various studies, specifically into the PHC practice setting (Lee et al., 2017). Many of the variables within assessment tools are not well adapted to a busy PHC environment (Hoogendijk et al., 2012). PCPs require the ability to identify frailty, but also manage the identified problems associated with that diagnosis (van Kempen, Melis, Perry, Schers, & Rikkert, 2015). Furthermore, the pace at which PCPs are having to see patients (usually in 5-15 minute slots) would also limit the usefulness of some of the lengthy assessments required to be able to screen for frailty, as some of the assessments have more than 40 variables within the tool (Lee et al., 2017). Lee et al. (2017) conclude that further research is needed to be able to identify specific frailty identifiers and recommend that assessments for use in PHC practice permit quick and efficient identification frailty (Lee et al., 2017).  1.3.3 Use of EMR for Frailty Research Advancement One promising method to identify frailty within PHC could be through using existing EMRs. EMRs are systems that enable health care professionals to record information gathered during a visit with a patient (Canada Health Infoway, 2018b). According to a recent survey, 81% 21  of Canadian PCPs are utilizing EMRs (Canadian Institute for Health Information, 2016). This is in part due to government funding, and recognition of the importance of electronic health initiatives by professional organizations like the Canadian Medical Association (CMA) (Chang & Gupta, 2015). The push for EMR adoption by health care providers is to be able to provide safe and efficient patient care, provide patient access to health information, and create meaningful use of the data that is inputted (Canadian Medical Association, 2014). Meaningful use of EMR data can be described as electronic capture of medical information in a coded way, which allows for tracking of key clinical conditions, and supports disease and medication management across the continuum of care (DesRoches & Miralles, 2011). Canada Health Infoway has many initiatives to help improve PHC EMR data and information across Canada (Canada Health Infoway, 2018a). This improvement through the use of content standardization, provides a framework for data to be utilized by researchers. This content standardization, also allows for gathering of larger data sets across provinces. With the increase in the use of EMRs by providers in Canada, and the subsequent increase in available patient data, we can start to develop ways we can use this information to monitor, treat, and improve patient outcomes (Zelmer & Hagens, 2014).  1.3.3.1 Canadian Primary Care Sentinel Surveillance Network (CPCSSN)  The Canadian Primary Care Sentinel Surveillance Network (CPCSSN) is the first pan-national database repository for chronic disease surveillance using de-identified EMR data from primary care practices (Birtwhistle et al., 2009; Williamson & Green, 2014). Because of the vast amount of data that can be pulled from EMRs, the administrators of CPCSSN have decided to focus on extracting and maintaining data for eight chronic illnesses (chronic obstructive pulmonary 22  disease [COPD], dementia, depression, diabetes, hypertension, osteoarthritis, parkinsonism, and epilepsy) that effect the Canadian PHC patient population (Williamson & Green, 2014). Because the CPCSSN extracts and monitors these eight chronic conditions and more recently, a few more, this can limit some of the usefulness of this data to apply to other conditions like frailty. In order for these data to be used nationally, the parameters for identifying frail patients would need to be added to the database architecture within CPCSSN. By mapping the defining characteristics of frailty, it allows other researchers internationally to map frailty parameters to their EMR data should they also be using SNOMED-CT. Thus, mapping between data elements that identify frailty are important for expanded research possibilities nationally and internationally.   1.4  Conclusion  Frailty is a condition affecting many Canadians; Frailty will continue to increase as the population ages. Being in a state of frailty increases the risks for hospitalization, mortality rates, and decreased state of health (Fried et al., 2001; Kenneth Rockwood & Mitnitski, 2011). While frailty research is increasing, there remains no widely agreed upon definition but an increasing number of different assessment tools. Few of these tools can easily be integrated into the PHC setting. Developing the capability to extract relevant data elements (e.g., laboratory results, prescribed medications, blood pressure, weight) out of EMRs in PHC for more meaningful use provides an opportunity to improve both assessment and research on frailty. Assessment, early identification and care planning in PHC for those who are frail may also improve the quality of patient care. 23  Chapter 2: Literature Review  2.1  Introduction This chapter describes existing frailty assessment tools which make up the defining characteristics of frailty. It discusses how frailty assessments could be integrated within EMRs and the important role of standardized health informatics language. This chapter introduces the availability of Canadian practice based research networks from which to draw large sets of data.  2.2 Literature Identification Relevant literature was identified through searches of the Medline and Cumulative Index of Nursing and Allied Health Literature (CINAHL) databases between January 2001- January 2018. Keyword search terms included: frail or frail elderly or frailty AND assessment OR diagnosis OR phenotype OR index OR definition. The following MeSH search terms were also used: Electronic Medical Record OR Electronic Health Record AND Primary Care or Family Practice AND Canada. Standardized terminology articles were identified using the keyword search terms: Standardized terminology AND SNOMED-CT or SNOMED OR Mapping. Further filters were applied to only show entries written in English. Relevant journal articles were then chosen based on their title and abstract for further exploration. After reading selected journal articles in full, the bibliographies were examined for additional research which may be relevant to the research questions. Additionally, authors who were identified as making major contributions to the subject of frailty were searched in the databases using the “author” search function for any other relevant research. Medical, Nursing, and Health Information 24  Technology texts were sought out as additional resources to provide information on the role of informatics.   2.3 Validated Frailty Assessment Measures A number of tools are available to assess frailty. The tools used for frailty identification should be valid and reliable, predict adverse clinical outcomes and according to some, predict patient responses outcomes to therapies and be supported by a biological causative theory (Clegg, Young, Iliffe, Rikkert, & Rockwood, 2013; Dent et al., 2016).  The tools used to assess for frailty use a variety of physical identifiers and classifiers to give a risk assessment score. Vergara et al., (2016) described how these tools fit into three different categories. The first category is identifying frailty based on the presence or absence of physical and cognitive abilities. The second category of tools use the clinical judgement of the practitioner to determine frailty. This includes having knowledge of the entire clinical picture of the patient and using that knowledge to ascertain the presence or absence of frailty. The third type of frailty tool includes identifying specified biomarkers that predispose or increase the risk of someone having frailty. These biomarkers are detected by using patient blood and DNA samples, although the validity of using this method for identifying frailty has yet to be fully determined. Finally, there are frailty tools being used in emerging research that include the use of standard lab data such as routine blood and urine tests to help identify frailty (Ritt, Jäger, Ritt, Sieber, & Gaßmann, 2017).  The most commonly used tools include the Comprehensive Geriatric Assessment (CGA), Frailty Index (FI), Frailty Phenotype and the Clinical Frailty Scale (CFS) (Fried et al., 2001; Jones, Song, & Rockwood, 2004; K. Rockwood, 2005; Kenneth Rockwood & Mitnitski, 2011). 25  These tools are discussed more in the following sections. Most of the assessments tools represent a continuum of frailty, from non-frail to severely frail (Dent et al., 2016). This allows the clinician to classify the degree to which a person is frail as well as allow for timely interventions to slow or reverse this condition.   2.3.1 Comprehensive Geriatric Assessment Widely considered to be the gold standard in assessing frailty, the comprehensive geriatric assessment (CGA) is cited as being one of the most relevant and accurate instruments available to determine the medical, functional, environmental and psychosocial deficits of older persons (Jones et al., 2004; van Kempen et al., 2015). The CGA produces an inventory of personalized health problems which can lead to the creation of a targeted care plan (Pilotto et al., 2017). The CGA varies in content depending on the context in which it is administered. It has been studied in a variety of clinical settings, showing good predictive validity for adverse clinical outcomes in both hospital and community dwelling patients (Pilotto et al., 2017). One of the main drawbacks to the CGA is the time required to administer this type of assessment. It is estimated that it takes upwards of 30 minutes to complete a CGA (Hamaker, Wildes, & Rostoft, 2017). Also, traditionally the CGA has been carried out by a physician or geriatrician. However, there is research currently under way to assess the accuracy, feasibility, and outcome factors associated with having general practitioners and registered nurses perform these assessments (Ferrat et al., 2018).  26  2.3.1 Frailty Index The FI is based on a theory that individuals become increasingly frail as they accumulate “deficits” (Mitnitski & Rockwood, 2014; Kenneth Rockwood & Mitnitski, 2011). A deficit is defined as a symptom, sign, disability, disease, and/or laboratory measurement which helps signify something that is “wrong” with a person (Rockwood & Mitnitski, 2007). The FI counts how many deficits a person has, rather than focusing on the specific nature of those deficits. The clinical nature of this assessment focuses on different criteria which account for the deficits in an individual. The FI assessment requires coding 40 variables, which include observing activities, assessing activities of daily living (ADL), impairments in cognition and physical performance, comorbid conditions, patient perception of health, and mood (Searle, Mitnitski, Gahbauer, Gill, & Rockwood, 2008). These variables are calculated into a numerical ratio (deficit score divided by number of variables), where a larger ratio indicates a greater severity of frailty (Rockwood & Mitnitski, 2011).   The FI has been well validated, and is highly cited in the literature (Buta et al., 2016; Dent et al., 2016). This FI has been applied to multiple data sets internationally, showing better predictability for adverse clinical outcomes in both community and hospital patients than other frailty assessment measures (Dent et al., 2016). The FI approach has been modified to a clinical model in mice, which is producing promising frailty intervention studies (Rockwood et al., 2017). While this method of frailty identification is robust, it remains time consuming and results are presented as a ratio, which may make it hard for clinicians who are unfamiliar with the tool to interpret the results. However, the FI can be derived from existing CGA data (Dent et al., 2016) and may provide for better long-term management and identification of worsening frailty 27  within an individual (Cesari et al., 2014), especially if it is used in the context of an EMR which could perform the mathematical calculation automatically.  2.3.2 Frailty Phenotype The frailty phenotype defines frailty as a syndrome that is identified when a patient presents with at least three out of five specific criteria: shrinking, weakness, poor endurance/energy, slowness and low physical activity level (Fried et al., 2001). It has been applied in multiple epidemiological studies, where it has shown good predictability of adverse clinical outcomes and mortality rates (Buta et al., 2016; Dent et al., 2016). The frailty phenotype does not necessarily require a robust examination of the patient, which may be useful in identifying individuals at risk within a practice. This tool lacks the ability to quantify the degree of frailty and specific care plan outcomes needed to monitor and treat frailty (Cesari et al., 2014). Furthermore, this tool lacks the psychosocial components associated with frailty and includes some assessments, like grip strength, which are not routinely done in PHC (Dent et al., 2016). The parameters included in the frailty phenotype assessment have often been modified by the researcher depending on the setting in which it is being administered. For example, the grip strength and gait assessments have been replaced with self-reporting measures. This modification impacts the predictive ability of this assessment (Theou, et al., 2015).   2.3.3 Clinical Frailty Scale The CFS is another widely used tool in the assessment of frailty in PHC as it is a brief way to categorize the severity of frailty (Rockwood, 2005). Originally developed by Rockwood (2005), this scale shows similar success rates in identifying frail populations compared to FI 28  (Theou et al., 2017). The CFS has been studied in a variety of clinical settings including hospital, geriatric, and primary care and is predictive of risk for hospitalization and death (Theou et al., 2017). This tool uses nine different criteria to rank individuals from being very fit to terminally ill. This tool is seen as a way to summarize a more comprehensive CGA and has been widely used and validated in different areas of medicine including the acute and community settings (Theou, Walston, et al., 2015). A benefit of this tool is that it is quick and simple to use, however it can be subjective, because the practitioner is required to make a clinical judgement without having to pick specific measurable parameters to produce the score.   2.4 Frailty Assessment in Primary Care The CGA has been mainly used in acute care settings and geriatrician practices. The FI, frailty phenotype and CFS have also been used, but mainly in acute care settings. Assessment of frailty in PHC is relatively nascent. In part, this is due to the amount of time needed to administer most of the tools. For example, there is information included in the CGA which may not be relevant to a diagnosis of frailty in PHC. Collecting historical employment and family composition information could be considered problematic for busy PHC settings, where time spent with patients is constrained.    Currently none of these tools are widely used to assess for frailty in PHC. While each of the discussed tools have their merits, it is important to select an assessment tool that is easily used, effective and not too time consuming. This will ensure that PCPs can integrate this type of assessment into their practice since most cases of frailty can be or are diagnosed in PHC settings where there is no geriatrician available.  29  2.4.1 CARES Model eCGA Some researchers have been conducting a frailty assessment based on CGA data  (Jones, Song, & Mitnitski, 2005; Jones et al., 2004). This frailty assessment is an adaptation the standard CGA, in combination with two validated frailty assessment tools to create a comprehensive frailty specific tool to use electronically in PHC (Garm, Park, & Song, 2018). The Community Action and Resources Empowering Seniors (CARES) model, as proposed by Garm et al. (2018) combine components of the FI and the CGA to produce what they call a Comprehensive Geriatric Assessment-Frailty Index (CGA-FI), which both have statistical merit in the diagnosis of frailty (Theou, Walston, et al., 2015).  The combination of these two tools are integrated into the Intrahealth EMR being used by PCPs in the Fraser Health Authority region of British Columbia (Garm et al., 2018). This electronic CGA (eCGA) is comprised of over 70 variables and was used as part of a comprehensive plan to measure and mitigate the effects of frailty in PHC (Theou et al., 2017). The pilot project provided education and support to PHC physicians and nurse practitioners to implement the eCGA in combination with other validated geriatric assessments to produce an FI score. The identified individuals were then provided with a wellness summary, FI score, and instructions on how to contact a free of charge telephone-based health coach to address some of the needs identified in the wellness plan. The individuals were then re-assessed six months later, where the eCGA was repeated to look for any improvement in FI scores (Theou et al., 2017). This initial pilot project showed a statistically significant decrease in the FI scores from time of assessment to the time of reassessment at 6 months (Garm et al., 2018; Theou et al., 2017) which represented an improvement in frailty levels. 30  Currently, work is being continued to implement this eCGA in additional municipalities within the Fraser Health Authority. It is hopeful that the addition of this frailty assessment tool embedded within the existing EMRs will allow for greater use across PHC and increase the ability for data collection.  2.5 Electronic Medical Record Use in Canadian Primary Health Care  One of the objectives of this paper is to explore how the existing frailty measurement tools can be used in combination with the data collected through the EMR in PHC. Therefore, it is important to understand the context in which EMRs are currently being used in Canadian PHC, as well as the role health informatics has in the ability to use these data. The literature available relating to EMR use in PHC in Canada is still relatively limited, and mainly focuses on research studies utilizing data from EMRs (Terry et al., 2012). Because EMR use within Canada continues to expand, it is important to understand how we can best utilize what is currently available to us to provide best practice for our patients. EMR use within Canadian medical practices has grown exponentially over the past decade. A recent survey shows that nearly 81% of PCPs are utilizing EMRs in their practice (Canadian Institute for Health Information, 2016). According to the survey, the majority of physicians are using their EMRs for a number of patient related activities including ordering of lab/diagnostics, accessing medications lists including drug interactions, and accessing hospital discharge records. Also, nearly 42% of respondents reported accessing both provincial patient information systems, as well as utilizing decision support tools (Canadian Medical Association, 2017). This means that a growing number of physicians are inputting large amounts of clinical data into information systems across the country. As Terry et al., (2012) found, in Canada there 31  is only a small amount of available literature using EMRs for research purposes. This provides opportunities of growth within the research. To date, to the author’s knowledge, data being input by PCPs have not been mapped to the defining characteristics of frailty as expressed through the validated frailty assessment. Because of the increase in uptake of EMR use within PHCs there have been a number of initiatives introduced at both the provincial and national level to support the meaningful use and adoption of this technology to support clinical practice and research. Funded by the Canadian government, Canada Health Infoway has invested large amounts of capital to increase the adoption and interoperability of EMR use across the country (Canada Health Infoway, 2018a). At the provincial level, Doctors of BC has introduced the Doctors Technology Office to provide ongoing clinical and technical support to all PCPs using EMR's in BC (Doctors of BC, 2017). There are also a number of research initiatives throughout the country to promote the increase in EMR-related research projects, with the aim to improve patient outcomes. One of the main challenges is the ability to meaningfully use the clinical data from EMRs. There is also a large selection of vendors which supply EMRs both nationally and globally, though there are little data currently available to quantify which vendors are being used. There are upwards of over 20 commonly used EMRs in use, as found by doing basic vendor search queries. The choice of EMR vendor is up to each individual PHC practice and their selection of these products reflects their financial investment, product features, and ease of product use (Skolnik, Timko, & Myers, 2011). Each of these software programs have different built in features, code and store data in different ways from one another. This makes it challenging for clinicians, researchers and other stakeholders to retrieve and use these data.  32  Currently, PCPs enter a variety of patient and administrative data into their EMR. The types of EMR data are historical, transactional, and transmitted; Historical data consist of demographic data (e.g., age, sex, ethnicity, height, weight) and family medical/surgical histories. Usually these types of data are entered once or occasionally updated. Transactional data are usually input at each visit, containing information such as: reason for visit, diagnosis, lab/diagnostic orders and results, and prescribed treatments (Venot & Cuggia, 2014). Lastly, EMRs can accept transmitted data from outside sources. These types of data would typically include hospital discharge reports, lab/diagnostic results, and specialist consultation reports (Venot & Cuggia, 2014). One of the other challenges of EMRs in PHC is related to the ability to transfer, and use data from one EMR system to another. This process of sharing the information between two or more systems, and use the data that has been exchanged, is referred to as interoperability (Nijeweme-d’Hollosy, van Velsen, Huygens, & Hermens, 2015). The EMR systems that are being used in PHC come from a variety of different vendors, which were not all designed to be interoperable with other systems. There are many different strategies in which to share information between these disparate systems. One of these strategies is to robustly develop standardized languages and terminologies that are understood by these different EMRs. Having standardized terminologies provides a pathway in which to reach interoperability (Dixon, Vreeman, & Grannis, 2014; Liyanage, Krause, & de Lusignan, 2015).  2.6 Standardized Terminologies To understand the requirements for extracting relevant data to use for decision support and/or research in relation to frailty, it is important to understand the role health informatics and 33  standardized terminology plays. A challenge in using EMR data is in part due to the various ways data can be entered into the EMR. EMR data are either unstructured (free text) or structured with discrete data. While free text entries provide the lowest barrier to adoption by the end user, structured data increases the ability for the data to be extracted and used in a more meaningful way (Cuggia, Avillach, & Daniel, 2014). Natural language through clinical descriptions entered by the provider, are the most common way of expressing information in an EMR, and are commonly found in free text notes such as clinical visit notes or discharge notes (Cuggia et al., 2014). While historically these types of data entry have been problematic in being able to extract data, there have been continued advancements in both Natural Language Processing and standardized languages which provide researchers the increased capability in meaningful data extraction (Cuggia et al., 2014). There has also been an increased amount of awareness for standardized languages which code clinical data in a uniform way, enabling for the extraction and automatic processing for the purposes of research.  In order for the information in an EMR to be used for research purposes, there need to be standardized ways in which to receive/communicate the information. There are a number of standardized language organizations within health informatics which produce standardized languages for the purpose of coding, sharing, and use of health information internationally (Cuggia et al., 2014). Standardized languages are used to help delineate different medical conditions and terminology used within health care. There are also standardized ways in which these systems communicate the information with one another across differing platforms. Because of this standardization, it makes it easier for researchers to use the inputted data across different countries and health care settings. 34  2.6.1 Standardized Clinical Terminologies There are many different standardized clinical terminologies currently in use within the healthcare domain. These clinical terminologies represent a need to be able to express clinical information in a systematic way, to enable communication of this information across heterogenous contexts, and be able to retrieve and store this information (Richesson & Krischer, 2007). Because human language is vast, and there are multiple ways of expressing complex clinical concepts and meanings, standardized language implementation remains complicated  (Andrews, Patrick, Richesson, Brown, & Krischer, 2008). Because human language expresses these concepts in different ways, these multiple concept expressions can be organized into a single representative term through a network of hierarchies. Each terminology domain has specific concept relationship and guiding principles which create the standardized terminologies (or nomenclature) which are specific to that domain (Hammond, Jaffe, Cimino, & Huff, 2014; Kim & Matney, 2014). There are two different types of terminologies which are present in EMR data entry. Interface terminologies are a systematic collection of specific health care terms, which aid the user in inputting and reading clinical information (Rosenbloom, Miller, Johnson, Elkin, & Brown, 2006). These terms are specific to the context and group in which they are being used, usually contain acronyms, and may change over time (Schulz, Rodrigues, Rector, & Chute, 2017). Reference terminologies are stable and well defined representational terms that coordinate with computer processable codes with aid in data aggregation and retrieval (Schulz et al., 2017)   35  2.6.1.1 Systematized Nomenclature of Medicine Clinical Term  Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) is an internationally used, multilingual, and most comprehensive clinical reference terminological system for medical nomenclature (SNOMED International, 2017). SNOMED CT is also an EMR interface terminology which can be used for clinical data entry, and represents clinical information and knowledge, and can help generate clinical decision support algorithms and retrievable clinical data (Bhattacharyya, 2016). This translates to a common language for describing patient problem lists, family histories and can be reproducible between providers. This standardized terminology is widely used, and the most common standard language currently in use in EMRs. The growth of SNOMED CT adoption has translated this as one of the preeminent standardized languages used for clinical and informatics research (Bhattacharyya, 2016)  Using SNOMED CT has both its benefits and challenges. Because it is a large repository, there are over 300,000 distinct clinical terms within it (SNOMED International, 2017). While this allows for great clinical concept coverage, it also leaves room for selection error and semantic ambiguity (Monsen et al., 2014). Having an expert clinician aid in the mapping process, as well as employing different mapping activities are some of the ways in which to ensure rigour and validity when using SNOMED CT (Block, 2016; Monsen et al., 2014).   2.6.1.2 Other Standardized Terminologies The International Classification of Diseases (ICD) is a classification system that is used internationally, maintained by the World Health Organization, and allows for the coding of all medical diagnoses and symptoms including those which are not otherwise specified (Cuggia et 36  al., 2014). This alpha numeric coding ensures that a medical diagnosis is understood across different languages and countries. Currently, PCPs use ICD codes to assign diagnoses for the purpose of data entry and administrative billing data. The Logical Observations Identifiers Names and Codes (LOINC) is a classification system used internationally for the standardized coding of laboratory and clinical observations using HL7 to send these messages between platforms (Regenstrief Institute, 2017). LOINC terminology is mainly used to represent measurement data, medical tests and results, whereas SNOMED CT focuses more on clinical terms, diagnoses, procedural codes, and some process codes. Standards organizations are in the process of linking LOINC terms with SNOMED CT to ensure broader coverage of standardized terminologies. Having both of these standardized terminologies integrated into an EMR system may ensure that there are fewer gaps in clinical content coverage.  2.6.2 Clinical Terminology Mapping Clinical terminology mapping is the process of linking a concept or code from one code set or system, with that of another code set or system with similar concept meaning (SNOMED International, 2017). The process of mapping concepts is important to be able to exchange information between various clinical systems for the purpose of data retrieval and use. This provides a means for otherwise non-interoperable systems to communicate. It also aids in the retrieval of data for analytical use by providing a defined record of computer readable coded clinical entities, which can be used in database architecture. The process of terminology mapping can happen through a variety of methods and requires clinical knowledge and scientific rigor to ensure effectiveness and reproducibility. There are special considerations and decisions that need 37  to be pre-determined in standardized terminology mapping for matching complex and specific clinical concepts. Clinical concepts can be mapped using “pre-coordination” and/or “post-coordination” methods. Pre-coordination clinical term matching, refers to having an exact match between the desired root clinical concept and the standardized terminology in which it is being mapped to. Post-coordination involves the combination of clinical concepts to satisfy the meaning of the root clinical concept. The researcher also needs to decide on which methods will be used to map the clinical concepts, and how they can achieve scientific rigor in those results (Coiera, 2015; Rosenbloom et al., 2006). The methods in which the author undertook the terminology mapping will be discussed in the next chapter.  2.6.3 Clinical Impact of Standardized Terminology Usage There are many advantages to having standardized terminologies within an EMR and/or research database. Benefits of having standardized terminologies in clinical practice include the facilitation of evidence-based practice through increased reliability and validity of collected data, and data mining opportunities to improve research and clinical decision making (Hardiker, Bakken, & Kim, 2011). Having coded concepts which relate to a specific disease, such as frailty, can allow for large data aggregation. This could provide researchers with the ability to identify and monitor specific parameters related to that disease within a large population. The coded list of frailty parameters could also provide researchers outside national databases information about frailty assessment and enable them to map frailty assessment to their specific EMR data.  Another possibility for using standardized terminologies within clinical practice is that it can provide a framework in which to create specific decision support tools (Bodenreider, 2008). These standardized terminologies would be linked to algorithms that allow for the detection and 38  identification of a problem when the applicable information is entered into the EMR, which would then alert the clinician to take action, or provide them with a link to evidence-based practice. In terms of frailty, the assessment information could be entered into the EMR questionnaire, it would then alert the clinician of the patient’s frailty score and provide them with real time support on how best to intervene for maximum outcome potential.  2.7 Summary In summary, frailty assessment through validated tools incorporated into an EMR used in PHC can be a valuable tool to assess for frailty. There are available tools currently being used in an electronic format in which to assess frailty. There are also databases which collect relevant EMR patient information available for utilization by researchers. However, to the author’s knowledge, frailty concepts have not been mapped to a structured EMR database. As a contribution to the existing and ongoing research in frailty within the PHC context, this paper seeks to map the components of this eCGA to standardized SNOMED CT codes. This will enable researchers to be able to add these data components to the CPCSSN database. This addition will allow researcher across the country to be able to extract defining characteristics of frailty to use for further research.   39  Chapter 3: Methods 3.1 Introduction This chapter outlines the methods used to map the data elements from the eCGA form to the SNOMED CT standardized terminology. This chapter provides an overview for the study design, sampling plan, procedure and analysis of the data. This research design and execution has been in collaboration with the research supervisor and advisor to ensure rigor and validity.  3.2 Research Questions This paper addresses two research questions: Research Question 1: What are the defining characteristics of frailty?  Are these characteristics captured in the eCGA form? Research Question 2: What data elements from an eCGA can be mapped to existing SNOMED CT standardized terminology and what is the rate of equivalence?   3.3 Study Design This study was conducted using a nonexperimental descriptive study design. This was completed by comparing the defining characteristics of frailty to the eCGA, then examining the rate of equivalence between the 133 predefined eCGA assessment data elements, with the corresponding SNOMED CT terms. The descriptive analysis was conducted in four phases; a) assessing whether the eCGA captures the defining characteristics of frailty as identified in the literature review; defining and comparing the characteristics of frailty with the eCGA form, b) creating a clinically relevant map of the eCGA form to help guide decision making in the mapping process, c) manual mapping of eCGA data elements to SNOMED CT; d) clinician 40  review of mapping results (Figure 3.1). Descriptive statistics were used to assess the rate of equivalence in which frailty can be identified using standardized terminology.    Figure 3.1 Overview of study method   3.4 Sampling Plan To complete the manual mapping activity, the 133 frailty defining data elements of the eCGA were used (Appendix A). The eCGA form was retrieved from the Intrahealth EMR platform which is currently being used throughout Fraser Health Authority, British Columbia. The original eCGA form used within the CARES model project comprised of 172 individual data elements. For the purpose of this paper, this number was reduced to 133, as these components are the only ones used to derive a frailty index score. The use of the data elements from this form were obtained with permission from the original CARES model researchers and were provided to the author in spreadsheet format.  41  3.5 Procedure and Data Collection  In the first phase of this study, I created a table of the defining characteristics of frailty, based on the review of the literature, and comparing that to the data elements within the eCGA form (Appendix A). The purpose of this was to ensure that the eCGA form captured the defining characteristics of frailty. The second phase of the study was to create a visual map to represent the eCGA form. The purpose of this map was to aid the author, by grouping the eCGA form data elements into clinically meaningful groupings. This aided in the manual mapping of the elements to SNOMED CT. The third phase of this study employed a manual mapping approach of the predefined eCGA form elements to SNOMED CT terminology. This manual mapping produced a list of corresponding SNOMED CT terms with a match criterion of; direct match, indirect match, or no match. In the fourth and final phase of this study, the author’s manual mapping was clinician reviewed to ensure accuracy and relevancy was appropriate.  3.6 Phase 1: The Defining Characteristics of Frailty Because the intention was to eventually integrate the data elements from the frailty eCGA form into a database, and there have been multiple ways to both define and assess for frailty, it was important to ensure that the eCGA form is capturing the defining characteristics of frailty as expressed in the current literature. The author first examined the recent relevant literature on frailty, as well as largely cited seminal works regarding frailty definitions, to make a list of the most common defining characteristics of frailty. This was done by creating a spreadsheet of the terms used to describe the characteristics of someone having frailty.   42  The author then created a visual table of the clinical assessment variables for frailty, which were retrieved from assessment tools described in Chapter 2 (see Table 4.1). This table includes the eCGA form which was mapped to SNOMED CT. The author then compared columns within the table to ensure that there were no missing variables from the eCGA form.   3.7 Phase 2: eCGA Data Sources and Concepts Map The user interface for the eCGA form is dense and cluttered, and not grouped into clinically relevant sections. To ensure that all form data elements were captured in a clinically meaningful way during the mapping procedures, and because SNOMED CT terminology is very granular, it was necessary to create a diagram model to visually display the concepts represented in the eCGA form. The creation of visual data models has been used, and shown to be helpful, in previous terminology mapping research (Block, 2016; Chow et al., 2015; Harris et al., 2015). Using a diagram map allows the researcher to visually relate the information from the form in a more clinically meaningful way, and it provides guidance to the researcher by giving context in decision making when matching clinical concepts from the root source to the standardized terminology.  The eCGA diagram model created for this paper visually depicts the different components of the eCGA form, and helps group data elements into clinically relevant categories to aid in decision making during the mapping process (Wade & Rosenbloom, 2008). The form is represented by the solid green box, with the solid dark blue boxes representing the components that derive the final frailty score. The exam data which accounts for the majority of the frailty score parameters, are then grouped into clinically meaningful groups (i.e., psychosocial, daily living) to aid the author in the mapping activity. This map was used to help guide the concept 43  and hierarchy category selection during the manual mapping process, as well as during the clinician validation process. The exam data variables, are most closely associated with the SNOMED CT hierarchy type of “findings”. The individual components of the eCGA form which calculate the frailty score are denoted with asterisks.   3.8 Phase 3: Manual Mapping to SNOMED CT Prior to the manual mapping activity, the author made some decisions related to SNOMED CT hierarchy type and placement of the clinical concepts in relation to SNOMED CT structure. Each SNOMED CT concept is organized into a hierarchy with preferred name and a unique numerical code. At the top of the SNOMED CT hierarchy is the root concept, to which the eCGA data elements are mapped. These concepts also contain descriptions which can contain many synonyms for the same concept. SNOMED CT relationships link like concepts to one another when their meanings are related to one another (is-a relationship) or part of a causal relationship (has-a relationship) (SNOMED International, 2017). Sub types of the root concept are referred to as ‘Top Level Concepts’ and form the major branches of the SNOMED CT hierarchy design. These top level concepts have dependent concepts, which branch out into more sub-type concepts. This type of organization leads to specific, granular, clinical concepts (SNOMED International, 2017). It was decided that all matched terms would be labelled as “pre-coordinated”. That is, a specified term will be an exact match between the desired root clinical concept and the standardized terminology to which it is being mapped. It was also decided that most of the chosen terms should fall within two top level concept SNOMED CT hierarchy types: findings and observable entity. In SNOMED CT the findings hierarchy represents the results from a 44  clinical observation, assessment or judgement. It also accounts for normal abnormal clinical states, and includes concepts used to represent diagnosis (SNOMED International, 2017). The observable entity hierarchy represents a question or assessment which can be performed and can produce a discrete answer. An example of this is blood pressure or gender (SNOMED International, 2017). The eCGA form data elements (n=133) were manually and independently mapped by the author (SB) using the latest edition of the international SNOMED CT opensource browser (SNOMED International, 2018). The SNOMED CT web browser is open source and free for use with a signed online user license agreement, which is automatically generated when you open the browser. Each eCGA specific data element was entered into the browser to obtain all possible results for that concept. The selected concept was then examined further for accuracy by examining the concept summary, diagram, and details within the SNOMED hierarchy. All data were entered into an excel spreadsheet with the eCGA root concept name, form responses, form codes, full SNOMED CT concept name, SNOMED CT hierarchy type, and unique identifying number. All possible matching results were entered into the spreadsheet and given colour coding (Hardiker et al., 2011) (Table 3.1 and Table 3.2).  Clinical concepts were entered into the browser with exact wording and filtered by hierarchy type (finding). A selected SNOMED CT code could fall within the observable entity class, if the clinician physically observes the client doing the action while in the office (e.g., timed get up and go test). Another rule that was followed during the mapping process was having the selected SNOMED CT code fall within the same Parent/Child category whenever possible. For eCGA questions which are assessing patient ‘impairment’, the selected SNOMED CT code should not delve into the severity of the impairment but fit into a yes/no answer (e.g., Impaired 45  hearing should not = moderate/severe/total hearing loss”). Finally, the selected SNOMED CT code should be descriptive of the question being asked and be reflective of what we are trying to assess in a frail patient. (e.g., Speech=talking, not Speech=communication) The author had to broaden the search using semantically similar descriptions for a clinical concept due to the complex nature of the SNOMED CT database. For example, when searching for terms such as ‘banking’, the term ‘money’ is also searched for relevant results. To further decision making in concept selection, the SNOMED CT concept had to satisfy a question/answer relationship to be selected, as this reflects the nature of the eCGA assessment. The clinical term was considered a match in the findings hierarchy if it matched the criteria of a “question” and an “answer”.  After completing the first version of the mapping, the author then re-entered the initial results into the SNOMED CT to refine the list of any multiple possible matches, based on feedback from the scholarly advisors and local database managers, to produce a final mapping spreadsheet.  Table 3.1 Clinical term mapping matching criteria Criteria Definition Direct Match The eCGA data element matches the clinical concept available in SNOMED CT. These matches are known as pre-coordinated.  Partial Match  One-to-many: The data element could be accurately described using more than one standardized clinical concept within SNOMED CT. (ie., data element “urinary incontinent” and SNOMED CT concepts “urinary incontinent” and “continent: dependent”) OR Partial Match: The data element didn’t directly match with a SNOMED CT code, but semantically is similar enough for the clinician to feel it could be appropriate (ie. Data element “cleaning independent” and SNOMED CT concept “able to tidy house”  No Match The eCGA data element does not have an equivalent clinical concept match in SNOMED CT  46  3.9 Phase 4: Clinician Consensus Mapping To complete the clinician consensus mapping and examine inter-rater reliability, it was decided that a subset of the 133 data elements be chosen for review and consensus. For this phase, two rounds of expert consensus mapping occurred to achieve the final results. In this final step, two sets of 10 randomized eCGA questions were given to a clinician to map independently. Both sets of eCGA questions were presented in an Excel spreadsheet for the clinician to fill out. The root data points to be mapped to SNOMED CT ranged from one to three responses. The columns in the spreadsheet related to SNOMED CT concepts were left blank for the clinician to fill out independently. The clinician was also provided with the SNOMED CT Starter Guide for reference.  The clinician doing the mapping activity, had experience in using the eCGA form for research purposes, but had no prior informatics, standardized terminology mapping or SNOMED CT knowledge. The clinician was not part of the original manual mapping, thus giving an independent review of the results. A brief overview was given by the author to help the clinician navigate the mapping activity.  After the initial set of eCGA questions was mapped by the clinician, the author, clinician and academic advisors met to examine the results. Inter-rater reliability scores were calculated for the first round of coding. A second round of clinician consensus mapping occurred to examine if inter-rated reliability between the clinician and the author’s selected clinical terms improved. For this process, a different set of randomly selected eCGA data elements were presented to the clinician in a new excel spreadsheet. The advisors provided supervision and guidance as the author and clinician discussed the results. 47  After the clinician mapping process, a spreadsheet was compiled to show the comparison between the authors and the clinician’s mapped results. In total, two different consensus mapping activities (n=20) were completed by the clinician subject matter expert. The criteria for matching between the author and the clinician are listed in Table 3.2.   Table 3.2 Clinician mapping term matching criteria Criteria Definition Direct Match The same SNOMED CT code was chosen for the eCGA data element. Partial Match A semantically similar code, or code from the same SNOMED CT hierarchy was chosen, but no direct match.  No Match There was no match between the author and clinician SNOMED CT code selection.   3.10 Data Analysis Data were analyzed using descriptive statistics to provide rates of equivalence between the eCGA data elements and SNOMED CT clinical terms. Rates of equivalence were calculated manually by the author. Consensus between the author and expert clinician was achieved by discussing the rationale and decision making during the selection of the SNOMED CT code, and coming to a mutually agreeable SNOMED CT code. Consensus on the data elements was also achieved through guidance from the supervising committee. A Kappa coefficient was calculated after both rounds of clinician consensus to assess how well the two observers classified the variables. The purpose of calculating after both rounds was to assess if there was any improvement after the first round.  48  Chapter 4: Results   There were four phases in this study. The first phase involved a literature review which identified the clinical characteristics of frailty. The second phase of the study involved creating a data sources and concept map to aid the author in subsequent phases. The third phase was a manual mapping of a frailty assessment to SNOMED-CT, and the fourth phase involved a clinician consensus mapping activity.  4.1 Phase 1: The Defining Characteristics of Frailty Keywords used to identify frailty were placed into a spreadsheet (Table 4.1). The defining characteristics of frailty fit into a number of physical, cognitive, and psychosocial deficit components. The key deficits associated with the concept of frailty were derived by making a list of deficits described in the frailty tools discussed in the literature review. The deficits associated with having frailty include weight loss, muscle weakness, physical exhaustion, physical slowness, low activity levels, cognitive impairment, low mood/attitude, incontinence, and co-morbid conditions and need for multiple medications.  Table 4.1 Defining Characteristics of Frailty Characteristic Searle et al., 2008 Rockwood,  2005 Fried et al.,  2001 Garm et al., 2017 Physical     Weight Loss x x x x Weakness x x x x Exhaustion x x x x Slowness x x x x Low Activity x x x x Incontinence x x x x Cognition     Cognitive Impairment x x  x Psychosocial    Low Mood/Attitude x x  x Co-morbidities x x  x Polypharmacy    x  49  Frailty has been described as being on a continuum (Cameron et al., 2015; Fried et al., 2001). It occurs on a gradient, with the first stage being classified as “pre-frail” or slightly frail. This is where a person is more susceptible to developing frailty, usually after a medical or physical “event” (Bandeen-roche et al., 2006). A person who is considered as pre-frail would have only a small number of the aforementioned deficits. At the end of the frailty continuum, a person would present with a large number of health deficits resulting in a failure to thrive, and ultimately death (Cameron et al., 2015).  The classification of frailty depends on the assessment tool being used and the way that tool quantifies frailty. Of the tools reviewed, only the eCGA FI captured all deficit components. Table 4.2 provides a visual comparison of assessment components in the frailty tools.   Table 4.2 Characteristics of Frailty in Assessment Tools Variable and Sub Category Frailty Index Clinical Frailty Scale Frailty Phenotype CGA-FI (CARES) Cognitive Status  x  x Montreal Cognitive Assessment    x Functional assessment staging of Alzheimer’s disease    x Dementia  x  x Psychosocial x   x Low Mood x   x Depression x   x Anxiety x   x Depression x   x    Energy/Outlook x x x x Fatigue  x x x Motivation    x Health Attitude x x  x Daytime Drowsiness  x  x Control of Life Events    x Other  x    Sleep    x    50  Table 4.2 (con’t) Characteristics of Frailty in Assessment Tools Variable and Sub Category Frailty Index Clinical Frailty Scale Frailty Phenotype CGA-FI (CARES) Physical Ability x x x x Vision    x Hearing    x Pain    x Functional reach    x Usual Activities   x x Exercise  x x x Strength x  x x Weakness Upper Proximal x   x Weakness Upper Distal x   x Weakness Lower Proximal    x Weakness Lower Distal    x    Mobility x x x x Balance    x Falls  x  x Walk Outside x x  x Walking x x x x Aid x x  x Stairs x    Timed get up and go x (pace)  x (timed 15 ft) x Weight x  x x Under   x x Over    x Obese    x Continence  x x  x Bowel  x  x Bladder  x  x ADL's x x  x Cooking x x  x Cleaning x x  x Shopping x x  x Medications x x  x Bed  x  x Driving  x  x Banking x x  x Dressing x   x Eating x   x Toilet x   x Bathing x x  x CoMorbidities x x  x # of Medications    x # of Problems x   x 51  4.2 Phase 2: eCGA Map of Data Sources and Concepts The eCGA Map of Data Sources and Concepts (see Figure 4.1) was created to help visualize the eCGA form and to group the components of the form into clinically meaningful categories to aid in the process of manual terminology mapping.  All parts of the eCGA form are expressed on this map. However, only the 133 components (see “*” in Figure 4.1, 4.2, and 4.3) which are used to derive a final frailty index score, were used to complete the terminology mapping activities. The diagram has been portioned into additional subsections for ease of readability (Figures 4.2 and 4.3), and a legend is provided below (Table 4.4).  Table 4.3 Legend-Overall Map of Data Sources and Concepts   Colour Description  Components of eCGA which derive Frailty score  Assessment Categories  Assessment variables  52  Figure 4.1 Overall Map of Data Sources and Concepts    53  Figure 4.2 Data Sources                 54  Figure 4.3 Exam Data   55  4.3 Phase 3: Manual Mapping to SNOMED CT All of the frailty assessment data elements were manually mapped using words from the eCGA form, using the mapping rules created by the author, as well as referring to the clinical map for guidance, and decision making related to clinical content and meaning.  With the criteria as described for manual mapping of the eCGA data elements, a total of 96/133 (72%) had a direct match, 22/133 (17%) had a partial match, and 15/133 (11%) had no match (see Appendix B for details).   Table 4.4 Manual Mapping of eCGA to SNOMED CT eCGA Data Elements (n=133) Direct Match One-to Many  No Match Total 96/133 22/133 15/133 Percentage 72% 17% 11%  4.4 Phase 4: Phase 4: Clinician Consensus Mapping For this phase, two rounds of expert consensus mapping occurred to achieve the final results. The initial set of eCGA questions was mapped by the clinician and the author. The initial consensus between the clinician and author was 60% direct match with one another, 10% partial match, and 30% no match (Table 4.4). A second round of clinician consensus mapping occurred to ensure there was enough inter-rated reliability between the clinician and the author’s selected clinical terms. This iteration had an 86% direct match with one another. In each iteration, there was 100% consensus on the mapped data elements after discussion and rationale were presented (see Table 4.5 and Appendices C and D).    56  Table 4.5 Clinician Consensus Mapping of eCGA to SNOMED CT- Iteration #1 eCGA Data Elements (n=20) Direct Match Partial Match  No Match Total 12/20 2/20 6/20 Percentage 60% 10% 30%  Table 4.6 Clinician Consensus Mapping of eCGA to SNOMED CT- Iteration #2 eCGA Data Elements (n=22) Direct Match Partial Match  No Match Total 19/22 None 3/22 Percentage 86%  14%  The Kappa values were initially weak (k= 0.33 for iteration #1) but were strong (k=0.75) for iteration #2. As some of the clinician consensus mapping informed the final list of SNOMED CT eCGA mapping codes, the final list as seen in Appendix B was comprised following this stage of the research.   4.5 Summary In this chapter, the defining characteristics of frailty, the eCGA clinical map, and results of manual and clinician consensus mapping were presented. Manual mapping results showed that 72% of the e-CGA mapped directly to SNOMED-CT, 17% were a partial match, and 11% of data elements had no match. The clinician consensus obtained a final result of an 86% direct match between the author and the clinician mapping and 14% no match. The author and clinician came to a 100% consensus on the selected terms, which produced the final list of mapped codes.   57  Chapter 5: Discussion 5.1 Introduction The work mapped the defining characteristics of frailty from an electronic assessment to standardized clinical terminology. We examined the rate of equivalence between the assessment and standardized terminology. The methods included comparing the defining characteristics of frailty to the data components in an eCGA, creating a visual map, and then manually mapping these components to SNOMED CT. After two rounds of clinician consensus mapping, the rate of equivalence and interrelated reliability were calculated. The final result was a list of the 133 eCGA components mapped to SNOMED CT codes.  This discussion includes the limitations and lessons learned and explores how this type of research could inform future standardized terminology activities to help advance frailty research. It also discusses the role of nursing within community-based care contexts and the nursing contribution to informatics and research.   5.2 Defining Frailty The defining characteristics of frailty fall into multiple deficit categories: physical, cognitive, and psychosocial. Clearly frailty, while considered a medical syndrome, is not easily identifiable with one type of diagnostic lab test or measurement (e.g., blood pressure). Key deficits associated with having frailty include weight loss, muscle weakness, physical exhaustion, physical slowness, low activity levels, cognitive impairment, low mood/attitude, incontinence, co-morbid conditions and the need for multiple medications (Fried et al., 2001; Jones, Song, & Rockwood, 2004;  Rockwood, 2005; Rockwood & Mitnitski, 2011). These characteristics are consistent with retrospective frailty EMR database research in primary care  58  patients (Anzaldi, Davison, Boyd, Leff, & Kharrazi, 2017). The Anzaldi et al. (2017) study found that patients identified as frail within the EMR also had the presence of a number of geriatric syndromes, such as incontinence, falls, and dementia, classified within the system. The complexity of assessing frailty has led to an abundance of tools and measures used to assess frailty (Dent et al., 2016). Of the five tools or measures reviewed in this work, the eCGA most comprehensively assessed the various deficit categories and therefore it was chosen to complete the mapping activity.  The defining characteristics of frailty and the visual representation of the frailty assessment variables were used to ensure the eCGA being used for the manual terminology mapping were encompassing of the aspects which would be needed for a PCP to assess and identify frailty in a patient. It was concluded that the eCGA, which is currently being used in frailty research projects, was the most complete of the tools assessed to identify frailty and frailty severity (Theou et al., 2017). The significance of having this eCGA as a comprehensive of frailty assessment is that this particular tool could be replicated within an EHR as a way for a PCP to be able to assess for frailty.  5.3 Implications of Defining Frailty Through the eCGA Tools such as the eCGA can be implemented within existing primary care EMRs in order to provide better frailty definition guidelines and assessment parameters. The eCGA-FI form does not appear to have any significant gaps that would prevent a clinician from accurately assessing for frailty. This could provide an accessible way for clinicians to identify patients who are frail. The goal of better frailty assessment is so that modifiable behaviours and early stage frailty interventions can be implemented in order to help prevent, slow, and/or reverse declines in  59  health status that are not directly related to gradual losses seen in normal aging. This model of implementing multi-component modifiable behaviours and interventions has been successful within primary care and community dwelling frail patients (Serra-Prat et al., 2017; Theou et al., 2017). These interventions include physical therapy, nutritional, and psychosocial components. At the end of the studies, patients had reduced their level of frailty, and had improved. By slowing the progression of frailty, clinicians are able to improve patients’ physical strength and thus their quality of life and hopefully prevent adverse outcomes such as falls, declining health, and hospitalizations (Gobbens, Luijikx, & Van Assen, 2013; Theou et al., 2017).   5.4 Manual Standardized Terminology Mapping This paper used a manual mapping technique to match the eCGA assessment elements to SNOMED CT terminology. The final manual mapping process produced a total of 133 uniquely matched SNOMED CT terms. This type of mapping method has previously been used in other nursing led standardized terminology mapping studies (Block, 2016; Hardiker et al., 2011). A number of other studies also incorporated an automated or semi-automated approach to terminology mapping. This increases rigour and ensures the greatest amount of clinical concept coverage within SNOMED CT (Block, 2016; Harris et al., 2015;  Kim et al., 2014; Monsen et al., 2014). However, due to the scope of this scholarly paper, it was decided not to incorporate an automated approach.  Previous studies have shown the importance that a clinically trained person’s perspective adds to the terminology mapping process (Kim, Kim, Shin, & Kim, 2012; Richesson, Andrews, & Krischer, 2006). In the Kim et al. (2012) study, it was discovered that there was a much higher accuracy of mapped terminology when the work is done by a clinician as opposed to a non- 60  trained clinician. In fact, the accuracy is even higher when the clinician is trained in the specific domain for which they are mapping (Kim et al., 2012). This shows the importance of having manual mapping being done by expert clinicians with knowledge of that specific domain. To satisfy this requirement, the terminology mapping was completed by the author who has terminology mapping and geriatric nursing experience. Furthermore, the expert clinician mapping was completed by a registered nurse who has experience in both assessing patients using the eCGA, but also teaching PCPs on how to use the eCGA tool.  One important aspect of the manual mapping process was the creation of a basic set of rules which the author and the clinician followed throughout the process. This rules-based system for terminology mapping has been used successfully in the past, and has been included in the recommendations of previous research (Kim et al., 2014; Wade & Rosenbloom, 2008). The creation of a set of rules is important, as it allows the terminology mapper to follow a set of guidelines to make decisions. This also increases rigour by allowing other people to use these same set of rules to complete similar mapping activities. These rules also allow for a more defined interpretation of the abundant concepts within SNOMED CT. As such, a rules-based approach can help the clinician select the most appropriate and semantically similar SNOMED CT code which they are trying to match.  5.4.1 Lessons Learned from Manual Mapping The first round of mapping and expert consensus was done without the creation of a rule set to follow to assist with decision making. This was, in part, due to the inexperience of the author in using SNOMED CT. After the rules were established, with the advisor’s guidance, a second round of mapping took place to refine the results.   61  The one to many matches, 17% (22/133) could be due to a lack of concept granularity or concept ambiguity both within the eCGA form and/or SNOMED CT clinical terms. An example of this is ambiguity are the eCGA questions regarding incontinence (e.g., Constipation-Yes or No). The author found two possible matches within SNOMED CT which would fit the criteria based on the rules established. The applicable answers in this case were “Infrequent Bowel Action (finding)” and “Decreased Frequency of Defecation (finding)”. While the majority of the mapped terms fell within the direct match category, it is possible that there could be more data elements which had one-to many and/or partial matches. This could be in part to the author having selection bias based on the rules established, and inexperience with mapping activities and using the SNOMED CT search engine.  The results obtained in this study show appropriate clinical concept coverage within SNOMED CT. Past work has shown manual mapping of concepts relevant to nursing and coverage in SNOMED CT to range from 58% to 83% (Harris et al., 2015, Ivory, 2016). The results obtained in this study suggest there is consistency between what was accomplished for frailty and previous work.  Some of the possible reasons for not having matches within SNOMED CT may be due to missing content within the terminology, but could also be due to the eCGA assessment form itself, and/or a lack of generalization to internationally recognized clinical terms. For example, the eCGA uses a timed sit-to-stand test, which, to the best of the author’s knowledge, is not part of the SNOMED CT database. It is possible that the timed sit-to-stand test is included in LOINC. Due to the scope of this paper, mapping to LOINC was not included. It is also possible that this content is missing, as it may not be used regularly enough for it to have been added to SNOMED CT. A third possibility for why a “no match” was found could be in part to the rules established  62  by the author. It is possible that there could have been a possible match which fell outside of these established rules. An example of this would be a matching term, but having it fall within a “disorder” or “procedure” hierarchy. Other studies in terminology mapping have similar explanations for the varying degrees of “no match”, or missing clinical content coverage in SNOMED CT. The degree to which content is missing, varies depending on the subject matter being mapped, the methods used to map, and the rules selected to complete the mapping activities (Harris et al., 2015; Ivory, 2016; Kim et al., 2014).   5.5 Expert Consensus Mapping  Two separate rounds of expert consensus mapping occurred in this study. Expert consensus mapping was chosen to increase the possible matches, as well as to determine inter rater reliability for the manual mapping completed by the author. The clinician taking part in this exercise had clinical expertise in the assessment of frailty, which are consistent with other studies which show the importance of specialized clinical knowledge to be able to map clinical concepts (Richesson et al., 2006; Wade & Rosenbloom, 2008).  The terms selected for the expert consensus mapping were randomly selected to increase reliability and rigour in the results. The first discussion and consensus between the two experts was facilitated with the help of the academic supervisor and advisor, to ensure proper procedure was followed. The first iteration of the expert consensus mapping shows a ‘fair’ amount of agreement between the two raters (McHugh, 2012a). This low level of agreement may be in part due to inexperience in terminology mapping by the second clinician. During the consensus conversation, with facilitation by the advisors, the two experts were able to discuss reasoning for concept selection. This discussion allowed for the discovery of the need for rules in the manual  63  mapping process. Because of the lack of agreement between the two raters, it was decided that the author should create a set of rules, in which to apply to the original list of manually mapped codes. Once these rules were applied, a second set of expert consensus mapping took place to ensure accuracy of results, once more rigour was applied to methods.  The second iteration of expert consensus mapping produced a ‘strong’ level of agreement suggesting that having a set of rules in which to base decisions on is important when completing manual mapping activities (McHugh, 2012b). It is also possible that if the second clinician had more prior experience in standardized terminology mapping, the inter related reliability would have been higher during the first iteration of consensus mapping.   In future it would be beneficial to have both clinicians be experts in the field related to the terminology mapping, and have prior mapping experience. It would therefore be beneficial, before undertaking terminology mapping, to provide basic training and education to clinicians related to standard terminologies and the process for mapping. This would ensure that the researchers have the same understanding of the terminology prior to mapping. A next step in this research would be to have a physician familiar with frailty assessment validate the list of mapped clinical terms in a similar fashion to the expert clinician consensus activity. A further possible direction of inquiry would be to have a separate mapping activity completed by a non-expert clinician, but expert terminology mapper to assess the differences between the two.   5.6 Implications The majority of the components found in a validated frailty assessment tool are present in SNOMED CT clinical concept terms. While only one method of terminology mapping was used, this research can help guide nursing-led informatics and standardized terminology mapping  64  activities by providing standard guidelines to follow in future endeavors. The most important implications coming from this research are related to education and data collection.   5.6.1  Clinical Workforce Education The data available from EMRs should be reliable and accurate to the type of clinical concepts they are trying to capture. Therefore, it is important to ensure that the clinical workforce of today be aware of standard terminologies and their application, as well as its importance to quality data collection. This is important for all PCPs including registered nurses, nurse practitioners and family practice physicians. To achieve this, it is important to have informatics education provided to nurses starting at the undergraduate level, with more specialized knowledge offerings at the graduate level (Darvish, Bahramnezhad, Keyhanian, & Navidhamidi, 2014). It is also important to have a standardized set of nursing informatics competencies, which would allow for consistency in education and advancement in this growing field of nursing. This knowledge would help prepare RNs entering the workforce in an increasingly electronic clinical environment. This would also help RNs understand the importance of clinical terminologies, and their usefulness in capturing nursing ‘work’, and contribute to quality data collection. The knowledge gained from learning standardized terminologies would provide a basis for RNs to share the same basic understanding, and could provide more consistent and higher quality data collection.  From a physician perspective, it would be beneficial if informatics education was part of medical training. However, this part of the curriculum has not been well established as part of standard medical education and training, with schools being reluctant to offer this education in an already saturated program. The benefit of including such education could potentially be a  65  better understanding of the flow of patient data, privacy and confidentiality issues, and how this data can be used in meaningful research (Shortliffe, 2010). Recent efforts have been made to study the incorporation of this type of education in medical training. Although there are many challenges to implementing this type of education, the institutions and students recognized the benefits this education provides (Sánchez-Mendiola, et al., 2012).  5.6.2 Data Collection Quality In an EMR it is important to ensure that data collection happens in a meaningful and standardized way to ensure complex concepts, like frailty, are accurately and consistently captured. Because concepts like frailty are dynamic and multidimensional, data elements stored in the EMR must be combined in a way to capture the complexities that conditions like frailty present. Using mapped standardized terminology to capture data elements can aid in this process. We want to ensure that these concepts are captured correctly, and oversight in EMR data organization for conditions like frailty is important, as accurate identification of frailty can lead to earlier intervention, and better care planning, which can increase the patient’s quality of life (Gobbens et al., 2013).   To ensure that these data are organized in a meaningful way, methods for concept mapping need to continue to evolve to ensure that quality and patient safety are achieved. If the data elements are not represented or mapped in an accurate way, an opportunity to collect meaningful data could be missed. Also, patient safety could be put at risk if these data are relied upon to inform care planning or decision support, as it may not accurately identify patients in the target population and prevent them from receiving proper treatment.   66   One way to mitigate these risks is by having nurses participate in more of these terminology mapping research activities. Nurses are ideal clinicians to further this research due to our educational and clinical assessment expertise. Nursing practice prepares nurses to have exceptional assessment skills in a multitude of areas. This allows nurses to better understand the context of clinical assessments and conditions for which they are trying to map to standardized terminology. Furthermore, an increasing number of nursing schools are including formal informatics training, including standard terminology, into their curriculum (Cummings, Borycki, & Madsen, 2015). This combined with basic understanding of research principles, makes RN’s ideal candidates to participate in standardized mapping activities   5.6.3 Mapping Research  Currently there is only a small amount of research available to draw from for this type of mapping. Part of the reason for this is because mapping can be seen as an operational procedure, rather than a research study. However, this type of mapping research can help show the importance of having clinicians be involved in this process of mapping and validation. This research varies in type of mapping, standardized terminologies, and clinical situations. This gap in research could be in part due to a lack of defined standards to guide other researchers. Also, this type of clinical concept mapping takes a considerable amount of time which may not be feasible from a cost or time perspective to organizations and institutions. The research that does exist mostly uses automated approaches to terminology mapping. While this is an important step that should be considered in any mapping activity, we should not discount the perspective a clinically trained person brings to the semantic meaning of the clinical concepts.    67  5.7 Limitations There are several limitations to this study. Due to the scope of this work, only one mapping method was used. To increase reliability and rigour, it would be advantageous to use additional mapping methods. One such method is an automated machine-based mapping process. This would enable exploration of a larger number of terms, and which might lead to different findings. It is possible that some of the results may be influenced by the author’s knowledge of both the manual mapping and expert consensus mapping activities. The author and the expert clinician were both RNs though neither specializes in geriatrics. This could have possibly affected the rate of equivalence and the final clinical term selection. This study could have been strengthened if more than two researchers completed mapping activities. Having a geriatrician or PCP familiar with CGAs would have helped to increase the strength.  With all terminology mapping there can be limited pragmatic use of the findings. One reason is that it is difficult to replicate the exact mapping methods, which can affect the validity of the findings (Saitwal et al., 2012). With manual mapping, the author selecting the terms may have some inherent selection bias which may differ from person to person. Also, because SNOMED CT has constant updates, upgrades, and different versions, it may make it difficult to replicate the results of this study at a later date. This limitation also highlights the challenge of adding this information into an EMR research database. Adding the final list of the SNOMED CT codes into a national data repository would require ongoing maintenance and sustainment to ensure that the codes selected are the most relevant and up to date. This could be problematic due to the time and cost related to this maintenance. The above limitations are not necessarily unique to this study, and should be noted for any future researcher wishing to undertake similar terminology mapping activities.   68   5.8 Recommendations As standardized terminologies continue to be integrated into EMRs, it is important to establish rules and parameters for researchers and clinicians to follow to ensure the meaningful use of these clinical terminologies. To strengthen the reliability of the mapped frailty assessment tool, this research should be peer reviewed and/or replicated by other researchers to validate the results. Ideally this should be done by a PCP who is familiar with performing CGAs for the purpose of identifying frailty. This would ensure that the clinical concepts selected are relevant to frailty assessment, and comprehensive of the identifying factors of frailty. It would also be beneficial if the validator had experience working with standardized terminologies such as SNOMED CT. However, if this is not feasible it would also be beneficial if the results from this research are reviewed by an informatics expert who is well versed in standardized terminology mapping. Having both an informatics and clinical expert review these findings would help ensure a high rate of reliability in the findings.  Currently, researchers and organizations are working to improve the use of standardized terminologies in EMRs to work towards greater interoperability between disparate systems, in part, by publishing inter-terminology mapping activities. To ensure that all clinical concepts are mapped, and to allow for greater interoperability between systems, it would be beneficial to also map the eCGA data elements to LOINC. Other studies have shown greater success in concept coverage, when data elements were mapped to both SNOMED CT and LOINC (Lougheed, Thomas, Wasilewski, Morra, & Minard, 2018). Because LOINC is more concerned with clinical measurement and laboratory data, it may be better at capturing some of the missing content from the eCGA. For example, there was “no match” available in SNOMED CT for the Montreal  69  Cognitive Assessment, or the 5 times sit-to-stand assessment. Both of these eCGA data elements are available within LOINC. The Regenstrief Institute (LOINC) and SNOMED International have an agreement to better integrate and harmonize the two terminologies to ensure less duplication through inter terminology mapping. This goal has just started to be realized, with the first edition of this cross mapping released to the public in January 2018 (SNOMED International, 2018).  Consistent with other terminology mapping studies, this research did not find complete concept coverage of the eCGA within SNOMED CT. This was in part due to missing content within SNOMED CT, specifically when assessing for the absence of a clinical problem. While the eCGA identifies frailty based on an accumulation of deficits, it is still an important step to record the absence of problems to be able to account for any changes to the baseline which could signify the onset or worsening of frailty. For example, there are questions regarding the presence or absence of a deficit (i.e., Delirium, Pain), or whether something is within normal limits (i.e., Speech). SNOMED CT is lacking the ability, in certain clinical concepts, to capture health normal baseline state, like absence of pain or delirium, but to capture that these concepts have been assessed.   5.8.1 CPCSSN Database and Recommendations  This study has been part of larger ongoing frailty research study being conducted by the academic supervisor. Part of this ongoing research involves adding a frailty case definition into a CPCSSN data repository (Birtwhistle et al., 2009; Williamson & Green, 2014). The CPCSSN collects data based on well-defined case definitions for each chronic illness. These case definitions were created using published evidence and expert physician guidance, using a  70  combination of ICD, numerical and textual data (Williamson & Green, 2014). The case definitions were validated using a robust methodology, and showed excellent to good sensitivity and validity for the 8 described illnesses (Williamson & Green, 2014).  However, these case definitions are discrete in nature. Meaning they function on a yes/no basis. Frailty for the purposes of identification and monitoring is more nuanced and on a continuum. Therefore, for the purpose of the CPCSSN database, it is important to clearly identify the characteristics of frailty. Because the CPCSSN currently has validated case definitions for eight chronic conditions, work is currently underway to use the EMR data for a case definition of frailty. By categorizing the defining characteristics of frailty, it allows us to identify the EMR data that could capture frailty. The recommendation after creating the case definition for frailty, would be to add the mapped standardized language to this database. Having the eCGA data elements mapped to SNOMED CT allows this work to be useful across Canada and internationally.   5.8.2 Canadian Primary Care Sentinel Surveillance Network Data The CPCSSN extracts almost all structured data (e.g. ICD-9, lab values, reason for visit, prescriptions) from community based primary care electronic medical records. These data form a large data repository which is updated with new data every six months. All data are stored on a secure network in a secure facility. No identifying information is extracted, except for postal code (Birtwhistle et al., 2009). Data are extracted from 17 different EMRs, and sit within a relational database in different tables including: encounters, demographics, prescribed medications, risk factor data such as height, weight, smoking status, coded health diagnosis, laboratory information and  71  physical examination data (Birtwhistle et al., 2009). CPCSSN data go through a three-step process that involves input into the CPCSSN database, data mapping to standard terms, and ‘cleaning’ of some data fields to make the information more usable for the end user (Birtwhistle et al., 2009). The data are loaded into a central repository where it is processed and added to the data warehouse (Birtwhistle et al., 2009). Currently, the CPCSSN database managers have proposed and accepted a recommendation to add an additional ‘measurement’ table into the data schema that would allow for the collection of eCGA data for frailty surveillance. However, the eCGA data are only coming from the Intrahealth EMR, there will need to be additional efforts to ensure that additional disparate EMR data can be integrated into this table. This would likely require some data preparation and cleaning from the other EMRs to ensure that it fits the data currently being collected. An outcome of this paper is a list of codified frailty characteristics, which can also be added as a column within the aforementioned data table. This provides a framework for the CPCSSN database managers, to map future EMR frailty data elements to. Having national data collection on frailty assessment parameters would give researchers the ability to track frailty trends nationally, and help to develop clinical decision support tools. The overall outcome goal of this being improving frailty assessment and treatment in primary care.   5.9 The Role of Nursing in PHC and Frailty Research Registered nurses have been vastly underutilized in Canadian primary care practices (Attwell, Rogers-Warnock, & Nemis-White, 2012; Kennedy, 2014; Oelke, Besner, & Carter, 2014). In part, this is due to the RN role not be well defined, leading to role ambiguity (Kennedy, 2014; Norful, Martsolf, Jacq, & Poghosyan, 2017). Registered nurses and nurse practitioners are  72  well equipped to be integral members of the interdisciplinary PHC team. In addition to the clinical services they offer to PHC patients, they also provide a holistic care approach which has been seen as beneficial by both patients and PCPs (Ehrlich, Kendall, & Muenchberger, 2012; Kennedy, 2014; Norful et al., 2017). The training and patient/provider experience that the RN brings to PHC, allows them to be able to build trusting relationships with their patients. This helps nurses to better understand the unique sub-populations of patients in PHC (Attwell et al., 2012). Because their practice also has a grass roots approach, RNs are often able to better understand and utilize community services which may improve overall health determinants through assessment, action, and patient education (Kennedy, 2014). Registered nurses have an opportunity to deliver invaluable services within PHC, including frailty screening and assessment. They are an optimal choice to initiate and manage chronic disease management in PHC (Lukewich, Edge, VanDenKerkhof, & Tranmer, 2014). Through their interactions with patients, RNs can more readily complete the full geriatric assessment, including any functional assessments such as the “sit to stand test”. Moreover, RNs’ relationship building and awareness in completing care planning, allows for a more integrated and holistic approach to frailty identification and management. Their relationships may help improve early frailty detection, and lead to improved intervention leading to better health outcomes.  Along with RN involvement in PHC, it is important that RNs continue to be involved in informatics and standardized terminology research. Organizations like the Canadian Nurses Association, Canadian Nursing Informatics Association, CIHI, and Canada Health Infoway have created an action plan to collect and use nursing data standards in Canada going forward (White, Nagle, & Hannah, 2016). It is imperative that informatics education be part of nursing education  73  at both the undergraduate and graduate levels, to ensure RNs are prepared to assist and understand these emerging terminologies and technologies.   5.10 Summary  This paper mapped the 133 eCGA data elements to SNOMED CT, after defining the characteristics of frailty through the available relevant literature and validated assessment tools. The main outcome of this paper is a list of codified frailty assessment data elements, that have been analyzed as having an exact, partial or no match to SNOMED CT terminology. These data elements have influenced the formation of a new table of data elements within the CPCSSN. While there was a good amount of concept coverage within SNOMED CT, this study highlights gaps in current standardized terminology databases. By sharing this research with agencies such as Canada Health Infoway and SNOMED International, it will address some of this missing content and concept ambiguity. It is with hope, that this research is shared and used by researchers wishing to surveil frailty through the CPCSSN database.   74  References Andrews, J. E., Patrick, T. B., Richesson, R. L., Brown, H., & Krischer, J. P. (2008). Comparing heterogeneous SNOMED CT coding of clinical research concepts by examining normalized expressions. Journal of Biomedical Informatics, 41(6), 1062–1069. https://doi.org/10.1016/j.jbi.2008.01.010 Anzaldi, L. J., Davison, A., Boyd, C. M., Leff, B., & Kharrazi, H. (2017). Comparing clinician descriptions of frailty and geriatric syndromes using electronic health records: a retrospective cohort study. BMC Geriatrics, 17(1), 248. https://doi.org/10.1186/s12877-017-0645-7 Attwell, D., Rogers-Warnock, L., & Nemis-White, J. (2012). Implementing Practice Management Strategies to Improve Patient Care: The EPIC Project. Healthcare Quarterly, 15(2), 46–51. Retrieved from https://www.longwoods.com/product/22913 Bandeen-roche, K., Xue, Q., Ferrucci, L., Walston, J., Guralnik, J. M., Chaves, P., … Fried, L. P. (2006). Phenotype of Frailty : Characterization in the Women ’ s Health and Aging Studies, 61(3), 262–266. Bhattacharyya, S. (2016). Introduction to SNOMED CT. Singapore: Springer. Birtwhistle, R., Keshavjee, K., Lambert-Lanning, A., Godwin, M., Greiver, M., Manca, D., & Lagace, C. (2009). Building a Pan-Canadian Primary Care Sentinel Surveillance Network: Initial Development and Moving Forward. The Journal of the American Board of Family Medicine, 22(4), 412–422. https://doi.org/10.3122/jabfm.2009.04.090081 Block, L. J. (2016). Mapping Wound Care Elements to SNOMED CT. University of British Columbia. Bodenreider, O. (2008). Biomedical Ontologies in Action : Role in Knowledge Management ,  75  Data Integration and Decision Support, 67–79. Bohnert, N., Chagnon, J., & Dion, P. (2015). Population projections for Canada (2013 to 2063), provinces and territories (2013 to 2038): Technical report on methodology and assumptions. Statistics Canada. Brown, R. T., & Covinsky, K. E. (2018). Frailty as an outcome in geriatrics research: Not ready for prime time? Annals of Internal Medicine, 168(5), 361–362. https://doi.org/10.7326/M17-3048 Buta, B. J., Walston, J. D., Godino, J. G., Park, M., Kalyani, R. R., Xue, Q., … Varadhan, R. (2016). Frailty assessment instruments : Systematic characterization of the uses and contexts of highly-cited instruments. Ageing Research Reviews, 26, 53–61. https://doi.org/10.1016/j.arr.2015.12.003 Cameron, I. D., Fairhall, N., Gill, L., Lockwood, K., Langron, C., Aggar, C., … Kurrle, S. (2015). Developing Interventions for Frailty, 2015. Canada Health Infoway. (2018a). Connected Health Information in Canada: A Benefits Evaluation Study, (April). Canada Health Infoway. (2018b). Electronic Medical Records. Retrieved from https://www.infoway-inforoute.ca/en/solutions/digital-health-foundation/electronic-medical-records Canadian Frailty Network. (2017a). Retrieved from http://old.cfn-nce.ca/media/575584/cfn-2016-17-annual-report.pdf Canadian Frailty Network. (2017b). Frailty Matters. Canadian Institute for Health Information. (2016). How Canada Compares: Results From The Commonwealth Fund 2015 International Health Policy Survey of Primary Care Physicians.  76  Retrieved from https://www.cihi.ca/sites/default/files/document/commonwealth_fund_2015_pdf_en.pdf Canadian Institute for Health Information. (2017). Primary Health Care. Canadian Medical Association. (2014). How can Canada achieve enhanced use of electronic medical records? Retrieved from https://legacy.cma.ca//Assets/assets-library/document/en/advocacy/Enhanced-Use-of-EMRs-Discussion-Paper-Final-May-2014.pdf Canadian Medical Association. (2016). The State of Seniors Healthcare in Canada. Retrieved from www.cma.ca/En/Lists/Medias/the-state-of-seniors-health-care-in-canada-september-2016.pdf Canadian Medical Association. (2017). Number and percent distribution of physicians by specialty and sex, Canada. Cesari, M., Calvani, R., & Marzetti, E. (2017a). Frailty in Older Persons. Clinics in Geriatric Medicine, 33(3), 293–303. https://doi.org/10.1016/j.cger.2017.02.002 Cesari, M., Calvani, R., & Marzetti, E. (2017b). Frailty in Older Persons. Clinics in Geriatric Medicine, 33(3), 293–303. https://doi.org/10.1016/j.cger.2017.02.002 Cesari, M., Gambassi, G., Van Kan, G. A., & Vellas, B. (2014). The frailty phenotype and the frailty index: Different instruments for different purposes. Age and Ageing, 43(1), 10–12. https://doi.org/10.1093/ageing/aft160 Chang, F., & Gupta, N. (2015). Progress in electronic medical record adoption in Canada Recherche Les progrès dans l ’ adoption du dossier médical électronique au Canada, 61, 1076–1084. Chow, M., Beene, M., O’Brien, A., Greim, P., Cromwell, T., DuLong, D., & Bedecarré, D.  77  (2015). A nursing information model process for interoperability. Journal of the American Medical Informatics Association, 22(3), 608–614. https://doi.org/10.1093/jamia/ocu026 Clegg, A., Young, J., Iliffe, S., Rikkert, M. O., & Rockwood, K. (2013). Frailty in elderly people. The Lancet, 381(9868), 752–762. https://doi.org/10.1016/S0140-6736(12)62167-9 Coiera, E. (2015). Guide to Health Informatics (3rd ed.). London: CRC Press. Cuggia, M., Avillach, P., & Daniel, C. (2014). Medical Informatics, e-Health. https://doi.org/10.1007/978-2-8178-0478-1 Cummings, E., Borycki, E. M., & Madsen, I. (2015). Teaching Nursing Informatics in Australia, Canada and Denmark. In CSHI. In Context Sensitive Health Informatics: Many Places, Many Users, Many Contexts, Many Uses (pp. 39–44). Darvish, A., Bahramnezhad, F., Keyhanian, S., & Navidhamidi, M. (2014). The Role of Nursing Informatics on Promoting Quality of Health Care and the Need for Appropriate Education, 6(6), 11–18. https://doi.org/10.5539/gjhs.v6n6p11 Dent, E., Kowal, P., & Hoogendijk, E. O. (2016). Frailty measurement in research and clinical practice: A review. European Journal of Internal Medicine, 31, 3–10. https://doi.org/10.1016/j.ejim.2016.03.007 DesRoches, C., & Miralles, P. (2011). Meaningful Use of Health Information Technology: What Does it Mean for Practicing Physicians? In Electronic Medical Records (pp. 1–14). New Jersey: Humana Press. Dixon, B. E., Vreeman, D. J., & Grannis, S. J. (2014). The long road to semantic interoperability in support of public health: experiences from two states. Journal of Biomedical Informatics, 49, 3–8. https://doi.org/10.1016/j.jbi.2014.03.011 Doctors of BC. (2017). Doctors Technology Office. Retrieved from  78  https://www.doctorsofbc.ca/doctors-technology-office Ehrlich, C., Kendall, E., & Muenchberger, H. (2012). Spanning boundaries and creating strong patient relationships to coordinate care are strategies used by experienced chronic condition care coordinators. Contemporary Nurse, 42(1), 67–75. https://doi.org/10.5172/conu.2012.42.1.67 Fedarko, N. S. (2011). The Biology of Ageing and Frailty. Clinical Geriatric Medicine, 27(1), 27–37. https://doi.org/10.1016/j.cger.2020.08.006 Ferrat, E., Bastuji-Garin, S., Paillaud, E., Caillet, P., Clerc, P., Moscova, L., & Audureau, E. (2018). Efficacy of nurse-led and general practitioner-led comprehensive geriatric assessment in primary care: protocol of a pragmatic three-arm cluster randomised controlled trial (CEpiA study). BMJ Open, 8(4). Fried, L. P., Ferrucci, L., Darer, J., Williamson, J. D., & Anderson, G. (2004). Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 59(3), 255–263. Fried, L , Ferrucci, L, Darer, J, Williamson, J , & Anderson, G. (2004). Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 59(3), M255–M263. Fried, L. P., Tangen, C. M., Walston, J., Newman, A. B., Hirsch, C., Gottdiener, J., … McBurnie, M. A. (2001). Frailty in Older Adults: Evidence for a Phenotype. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 56(3), M146–M157. https://doi.org/10.1093/gerona/56.3.M146  79  Garm, A., H Park, G., & Song, X. (2018). Using an Electronic Comprehensive Geriatric Assessment and Health Coaching to Prevent Frailty in Primary Care: The CARES Model. Medical & Clinical Reviews, 1–5. https://doi.org/10.1088/0370-1298/69/1/308 Gobbens, R., Luijikx, K., & Van Assen, M. (2013). Explaining quality of life of older people in the Netherlands using a multidimensional assessment of frailty. Quality of Life Research, 22(8), 2051–2061. https://doi.org/10.1007/s11136-012-0341-1 Hamaker, M. E., Wildes, T. M., & Rostoft, S. (2017). Time to stop saying geriatric assessment is too time consuming. Journal of Clinical Oncology, 35(25), 2871–2874. https://doi.org/doi:10.1200/JCO.2017.72.8170 Hammond, W. E., Jaffe, C., Cimino, J. J., & Huff, S. M. (2014). Standards in biomedical informatics. In Biomedical informatics (pp. 211–253). Springer, London. Hardiker, N. R., Bakken, S., & Kim, T. Y. (2011). Advanced terminological approaches in nursing. In K. A. M. K. Saba (Ed.), Essentials of nursing informatics (5th ed.). New York: McGraw-Hill Medical. Harmand, M., Meillon, C., Bergua, V., Tabue Teguo, M., Dartigues, J. F., Avila-Funes, J. A., & Amieva, H. (2017). Comparing the predictive value of three definitions of frailty: Results from the Three-City study. Archives of Gerontology and Geriatrics, 72(April 2016), 153–163. https://doi.org/10.1016/j.archger.2017.06.005 Harris, M. R., Langford, L. H., Miller, H., Hook, M., Dykes, P. C., & Matney, S. A. (2015). Harmonizing and extending standards from a domain-specific and bottom-up approach: an example from development through use in clinical applications. Journal of the American Medical Informatics Association. https://doi.org/10.1093/jamia/ocu020 Hoogendijk, E. O., Van Der Horst, H. E., Deeg, D. J., Frijters, D. H., Prins, B. A., Jansen, A. P.,  80  ..., & Van Hout, H. P. (2012). The identification of frail older adults in primary care : comparing the accuracy of five simple instruments, (October 2012), 262–265. https://doi.org/10.1093/ageing/afs163 Ivory, C. H. (2016). Mapping perinatal nursing process measurement concepts to standard terminologies. , 34(7), 312. Computers, Informatics, Nursing: CIN, 34(7), 312–320. Jones D, Song X, Mitnitski A, R. K. (2005). Evaluation of a frailty index based on a comprehensive geriatric assessment in a population based study of elderly Canadians. Aging Clinical and Experimental Research, 17, 465–471. Jones, D. M., Song, X., & Rockwood, K. (2004). Operationalizing a Frailty Index from a Standardized Comprehensive Geriatric Assessment. Journal of the American Geriatrics Society, 52(11), 1929–1933. https://doi.org/10.1111/j.1532-5415.2004.52521.x Kennedy, V. (2014). The Value of Registered Nurses in Collaborative Family Practice : Enhancing Primary Healthcare in Canada. Nursing Leadership, 27(1), 32–44. Kim, T. Y., Matney, S. A. (2014). Standards. In R. Nelson & N. Staggers (Ed.), Health informatics: An interprofessional approach (pp. 351–369). St. Louis: Elsevier Mosby. Kim, S. Y., Kim, H. H., Shin, K. H., & Kim, H. S. (2012). Comparison of Knowledge Levels Required for SNOMED CT Coding of Diagnosis and Operation Names in Clinical Records, 18(3), 186–190. Kim, T. Y., Hardiker, N., & Coenen, A. (2014). Inter-terminology mapping of nursing problems. Journal of Biomedical Informatics, 49, 213–220. https://doi.org/10.1016/j.jbi.2014.03.001 Lee, L., Patel, T., Hillier, L. M., Maulkhan, N., Slonim, K., & Costa, A. (2017). Identifying frailty in primary care: A systematic review. Geriatrics and Gerontology International, 17(10), 1358–1377. https://doi.org/10.1111/ggi.12955  81  Liyanage, H., Krause, P., & de Lusignan, S. (2015). Using ontologies to improve semantic interoperability in health data. Journal of Innovation in Health Informatics, 22(2), 309–315. https://doi.org/10.14236/jhi.v22i2.159 Lougheed, M. D., Thomas, N. J., Wasilewski, N. V., Morra, A. H., & Minard, J. P. (2018). Use of SNOMED CT® and LOINC® to standardize terminology for primary care asthma electronic health records. Journal of Asthma, 55(6), 629–639. https://doi.org/10.1080/02770903.2017.1362424 Lukewich, J., Edge, D. S., VanDenKerkhof, E., & Tranmer, J. (2014). Nursing Contributions to Chronic Disease Management in Primary Care, 44(2), 103–110. https://doi.org/10.1097/NNA.0000000000000033 McHugh, M. L. (2012a). Interrater reliability: the kappa statistic. Biochemia Medica: Biochemia Medica, 22(3), 276–282. McHugh, M. L. (2012b). Interrater reliability: the kappa statistic. Biochemia Medica, 22(3), 276–282. Ministry of Health. (2017). Frailty in Adults- Early Identification and Managment. Retrieved from https://www2.gov.bc.ca/assets/gov/health/practitioner-pro/bc-guidelines/frailty-full_guideline.pdf Mitnitski, A., & Rockwood, K. (2014). Aging as a Process of Deficit Accumulation: Its Utility and Origin (Vol. 2015, pp. 85–98). https://doi.org/10.1159/000364933 Monsen, K. A., Finn, R. S., Fleming, T. E., Garner, E. J., LaValla, A. J., & Riemer, J. G. (2014). Rigor in electronic health record knowledge representation: lessons learned from a SNOMED CT clinical content encoding exercise. Informatics for Health and Social Care, 1–15. https://doi.org/10.3109/17538157.2014.965302  82  Nijeweme-d’Hollosy, W., van Velsen, L., Huygens, M., & Hermens, H. (2015). Requirements for and Barriers towards Interoperable eHealth Technology in Primary Care. IEEE Internet Computing, 19(4), 10–19. https://doi.org/10.1109/MIC.2015.53 Norful, A., Martsolf, G., Jacq, K. De, & Poghosyan, L. (2017). Utilization of registered nurses in primary care teams : A systematic review. International Journal of Nursing Studies, 74(November 2016), 15–23. https://doi.org/10.1016/j.ijnurstu.2017.05.013 Oelke, N. D., Besner, J., & Carter, R. (2014). The evolving role of nurses in primary care medical settings. International Journal of Nursing Practice, 20(6), 629–635. https://doi.org/10.1111/ijn.12219 Pilotto, A., Cella, A., Pilotto, A., Daragjati, J., Veronese, N., Musacchio, C., … Prete, C. (2017). Three Decades of Comprehensive Geriatric Assessment : Evidence Coming From Different Healthcare Settings and Speci fi c Clinical Conditions. Journal of the American Medical Directors Association, 18(2), 192.e1-192.e11. https://doi.org/10.1016/j.jamda.2016.11.004 Regenstrief Institute. (2017). What LOINC Is. Retrieved from https://loinc.org/get-started/what-loinc-is/ Richesson, R. L., Andrews, J. E., & Krischer, J. P. (2006). Use of SNOMED CT to Represent Clinical Research Data: A Semantic Characterization of Data Items on Case Report Forms in Vasculitis Research. Journal of the American Medical Informatics Association, 13(5), 536–546. https://doi.org/https://doi.org/10.1197/jamia.M2093 Richesson, R. L., & Krischer, J. (2007). Data Standards in Clinical Research: Gaps, Overlaps, Challenges and Future Directions. Journal of the American Medical Informatics Association, 14(6), 687–696. https://doi.org/10.1197/jamia.M2470 Ritt, M., Jäger, J., Ritt, J. I., Sieber, C. C., & Gaßmann, K. G. (2017). Operationalizing a frailty  83  index using routine blood and urine tests. Clinical Interventions in Aging, 12, 1029–1040. https://doi.org/10.2147/CIA.S131987 Rockwood, K. (2005). A global clinical measure of fitness and frailty in elderly people. Canadian Medical Association Journal, 173(5), 489–495. https://doi.org/10.1503/cmaj.050051 Rockwood, K., Blodgett, J. M., Theou, O., Sun, M. H., Feridooni, H. A., Mitnitski, A., … Howlett, S. E. (2017). A Frailty Index Based On Deficit Accumulation Quantifies Mortality Risk in Humans and in Mice. Scientific Reports, 7(1), 43068. https://doi.org/10.1038/srep43068 Rockwood, K., & Mitnitski, A. (2007). Frailty in Relation to the Accumulation of Deficits. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 62(7), 722–727. https://doi.org/10.1093/gerona/62.7.722 Rockwood, K., & Mitnitski, A. (2011). Frailty Defined by Deficit Accumulation and Geriatric Medicine Defined by Frailty. Clinics in Geriatric Medicine, 27(1), 17–26. https://doi.org/10.1016/j.cger.2010.08.008 Rodríguez-Mañas, L., Féart, C., Mann, G., Viña, J., Chatterji, S., Chodzko-Zajko, W., … Vega, E. (2013). Searching for an operational definition of frailty: A delphi method based consensus statement. the frailty operative definition-consensus conference project. Journals of Gerontology - Series A Biological Sciences and Medical Sciences, 68(1), 62–67. https://doi.org/10.1093/gerona/gls119 Romera-Liebana, L., Orfila, F., Segura, J. M., Real, J., Fabra, M. L., Möller, M., … Foz, G. (2018). Effects of a Primary Care-Based Multifactorial Intervention on Physical and Cognitive Function in Frail, Elderly Individuals: A Randomized Controlled Trial. The  84  Journals of Gerontology: Series A, 00(00), 1–7. https://doi.org/10.1093/gerona/glx259 Rosenbloom, S. T., Miller, R. A., Johnson, K. B., Elkin, P. L., & Brown, S. H. (2006). Interface Terminologies: Facilitating Direct Entry of Clinical Data into Electronic Health Record Systems. Journal of the American Medical Informatics Association, 13(3), 277–288. https://doi.org/10.1197/jamia.M1957 Saitwal, H., Qing, D., Jones, S., Bernstam, E. V., Chute, C. G., & Johnson, T. R. (2012). Cross-terminology mapping challenges: a demonstration using medication terminological systems. Journal of Biomedical Informatics, 45(4), 613–625. Sánchez-Mendiola, M. Martínez-Franco, A. I. Rosales-Vega, A., Villamar-Chulin, J., Gatica-Lara, F., García-Durán, R., & Martínez-González, A. (2012). Development and implementation of a biomedical informatics course for medical students: challenges of a large-scale blended-learning program. Journal of the American Medical Informatics Association, 20(2), 381–387. https://doi.org/10.1136/amiajnl-2011-000796 Schulz, S., Rodrigues, J. M., Rector, A., & Chute, C. G. (2017). Interface Terminologies, Reference Terminologies and Aggregation Terminologies: A Strategy for Better Integration. Studies in Health Technology and Informatics, 245, 940–944. https://doi.org/10.3233/978-1-61499-830-3-940 Searle, S. D., Mitnitski, A., Gahbauer, E. A., Gill, T. M., & Rockwood, K. (2008). A standard procedure for creating a frailty index. BMC Geriatrics, 8(1), 24. https://doi.org/10.1186/1471-2318-8-24 Serra-Prat, M., Sist, X., Domenich, R., Jurado, L., Saiz, A., Roces, A., … Papiol, M. (2017). Effectiveness of an intervention to prevent frailty in pre-frail community-dwelling older people consulting in primary care : a randomised controlled trial, (January), 401–407.  85  https://doi.org/10.1093/ageing/afw242 Shortliffe, E. (2010). Biomedical informatics in the education of physicians. JAMA: The Journal of the American Medical Association, 304(11), 1227–1228. Skolnik, N. S., Timko, M., & Myers, C. (2011). A view from the trenches: primary care physicians on electronic health records. In Electronic Medical Records (pp. 15–36). Totowa, NJ.: Humana Press. https://doi.org/10.1007/978-1-60761-606-1 SNOMED International. (2017). 5-Step Briefing. Retrieved August 20, 2001, from https://www.snomed.org/snomed-ct/five-step-briefing SNOMED International. (2018). SNOMED CT Browser. Retrieved August 20, 2008, from https://browser.ihtsdotools.org/? Terry, A. L., Stewart, M., Fortin, M., Wong, S. T., Kennedy, M., Burge, F., … Thind, A. (2012). How does Canada stack up? A bibliometric analysis of the primary healthcare electronic medical record literature. Informatics in Primary Care, 20(4), 233–240. https://doi.org/10.14236/jhi.v20i4.2 Theou, O., Cann, L., Blodgett, J., Wallace, L. M. K., Brothers, T. D., & Rockwood, K. (2015). Modifications to the frailty phenotype criteria: Systematic review of the current literature and investigation of 262 frailty phenotypes in the Survey of Health, Ageing, and Retirement in Europe. Ageing Research Reviews, 21, 78–94. https://doi.org/10.1016/j.arr.2015.04.001 Theou, O., Park, G. H., Garm, A., Song, X., Clarke, B., & Rockwood, K. (2017). Reversing Frailty Levels in Primary Care Using the CARES Model. Canadian Geriatrics Journal, 20(3), 105–111. https://doi.org/10.5770/cgj.20.274 Theou, O., Rockwood, M. R. H., Mitnitski, A., & Rockwood, K. (2012). Disability and co-morbidity in relation to frailty : How much do they overlap ? Archives of Gerontology and  86  Geriatrics, 55(2), e1–e8. https://doi.org/10.1016/j.archger.2012.03.001 Theou, O., Walston, J., & Rockwood, K. (2015). Operationalizing Frailty Using the Frailty Phenotype and Deficit Accumulation Approaches. Interdisciplinary Topics in Gerontology and Geriatrics, 41, 66–73. https://doi.org/10.1159/000381164 van Kempen, J. A. L., Melis, R. J. F., Perry, M., Schers, H. J., & Rikkert, M. G. M. O. (2015). Diagnosis of Frailty after a Comprehensive Geriatric Assessment: Differences between Family Physicians and Geriatricians. The Journal of the American Board of Family Medicine, 28(2), 240–248. https://doi.org/10.3122/jabfm.2015.02.130081 Vanselow, N. A. (1995). A New Definition of Primary Care. JAMA: The Journal of the American Medical Association, 273(3), 192. https://doi.org/10.1001/jama.1995.03520270026023 Venot, A., & Cuggia, M. (2014). Medical Informatics, e-Health, 349–363. https://doi.org/10.1007/978-2-8178-0478-1 Vergara, I., Rivas-Ruiz, F., Vrotsou, K., Contreras-Fernández, E., Téllez-Santana, T., MacHón, M., … Abellan, G. (2016). Validation and comparison of instruments to identify frail patientes in primary care settings: Study protocol. BMC Health Services Research, 16(1), 1–7. https://doi.org/10.1186/s12913-016-1540-1 Wade, G., & Rosenbloom, S. T. (2008). Experiences mapping a legacy interface terminology to SNOMED CT. BMC Medical Informatics and Decision Making, 8(Suppl 1), S3. https://doi.org/10.1186/1472-6947-8-S1-S3 White, P., Nagle, L., & Hannah, K. (2016). Strategizing for national nursing data standards. Retrieved November 28, 2018, from https://www.canadian-nurse.com/en/articles/issues/2016/november-2016/strategizing-for-national-nursing-data- 87  standards Williamson, T., & Green, M. E. (2014). Validating the 8 CPCSSN Case Definitions for Chronic Disease Surveillance in a Primary Care Database of Elec- tronic Health Records, 367–373. https://doi.org/10.1370/afm.1644.Department Woods, N. F., Lacroix, Ã. A. Z., Gray, S. L., Pharm, D., Aragaki, A., Cochrane, B. B., … Ãã, M. P. H. (2005). Frailty : Emergence and Consequences in Women Aged 65 and Older in the Women ’ s Health Initiative Observational Study, 1321–1330. https://doi.org/10.1111/j.1532-5415.2005.53405.x Xue, Q. L. (2011). The frailty syndrome: definition and natural history. Clinics in Geriatric Medicine, 27(1), 1–15. Zelmer, J., & Hagens, S. (2014). Advancing Primary Care Use of Electronic Medical Records in Canada. Health Reform Observer - Observatoire Des Réformes de Santé, 2(3). https://doi.org/10.13162/hro-ors.v2i3.1214   88  Appendices Appendix A   CARES CGA Questionnaire    89  Appendix B  Manual Mapping The tables below shows phase 3 mapping of eCGA data elements to SNOMED CT terms.  Criteria Definition Direct Match The eCGA data element matches the clinical concept available in SNOMED CT. These matches are known as pre-coordinated.  Partial Match  One-to-many: The data element could be accurately described using more than one standardized clinical concept within SNOMED CT. (ie., data element “urinary incontinent” and SNOMED CT concepts “urinary incontinent” and “continent: dependent”) OR Partial Match: The data element didn’t directly match with a SNOMED CT code, but semantically is similar enough for the clinician to feel it could be appropriate (ie. Data element “cleaning independent” and SNOMED CT concept “able to tidy house” No Match The eCGA data element does not have an equivalent clinical concept match in SNOMED CT     Form Element Description Form Code Response SNOMED Name SNOMED Heirarchy Type SNOMED Code 1 CGA Anxiety No  Normal emotional state   Finding 409068001     Yes Anxiety  Finding 48694002 2 CGA Balance Within Normal Limits  Balance normal  Finding 298312007     Impaired  Impairment of balance  Finding 387603000 3 CGA Cognition Within Normal Limits Normal cognition Finding 449888003    CIND/MCI Minimal cognitive impairment  Finding 110352000    Dementia Severe cognitive impairment Dementia Finding Disorder 702956004 52448006   4 CGA Control of Life Events Yes Maintains self-control  Finding 284475009     No Low self-control Finding 705000008 5 CGA Delirium Yes Delirious Finding 419567006     No No Match     6 CGA Delusion Yes Delusions Finding 2073000     No No Match     7 CGA Depression Yes Symptoms of depression  Depressive disorder  Finding Disorder    90      No No Match     8 CGA Daytime Drowsiness Yes Daytime somnolence Finding 141000119100     No       9 CGA Elimination Bladder Continent Continence independent   Bladder: fully continent  Finding 129023004 165234001    Incontinent Urinary incontinence Continence dependent Finding 165232002 129077000 10 CGA Elimination Bowel Continent  Bowels: fully continent  Finding 24029004     Incontinent Incontinence of feces Finding 72042002 11 CGA Elimination Catheter Yes Urinary catheter in situ Finding 439053001     No No Match     12 CGA Elimination Constipation Yes  Infrequent Bowel Action Decreased frequency of defecation  Finding  249516000 44316003     No Not constipated Finding 162081000 13 CGA Emotional Other   No Match     14 CGA Enough Income Yes Income sufficient to meet needs Finding 224190007     No Income insufficient to meet needs Finding 224191006 15 CGA Exercise Frequent Exercises regularly Finding 228448000    Occasional Gets little exercise Finding 228446001    Not Gets no exercise Finding 228445002 16 CGA Fall Yes  Falls Finding 161898004     No Does not fall Finding 298345007 17 CGA Fall Number Free Text Number of falls Observable Entity 298348009 18 CGA FAST Score   Functional status index  Assessment Scale 273472005 19 CGA Fatigue Yes Able to sustain energy level Finding 716453008     No Fatigue Finding 84229001 20 CGA Hallucination Yes Hallucinations Finding 7011001     No Normal perception Finding 247700009 21 CGA Health Attitude Excellent General health excellent  Finding 135816001    Good General health good  Finding 135815002    Fair General health fair Finding 135817005    Poor General health poo Finding 135818000    Pt Couldn’t Say Not sure of general health  Finding 135820002 22 CGA Hearing Within Normal Limits Hearing normal Finding 162339002    Impaired Hearing problem Finding 300228004  91  23 CGA IALDs Banking Independent Money managing independent Finding 129054000    Assisted   Money managing assisted Finding 129028008    Dependent Money managing dependent  Finding 129066006 24 CGA IALDs Cleaning Independent Able to tidy house Finding 286185000    Assisted Needs help with housework Finding 400985006    Dependent Unable to tidy house Finding 286186004 25 CGA IALDs Cooking Independent Independent in cooking  Finding 710763006    Assisted Needs help with cooking Finding 40986007    Dependent Unable to cook food Finding 286515001 26 CGA IALDs Driving Independent Able to drive a car Finding 300634001    Assisted Difficulty driving a car Finding 300638003    Dependent Unable to drive a car Finding 300635000 27 CGA IALDs Meds Independent Able to manage medication Finding 285034004    Assisted  Difficulty managing medication   Finding 285038001    Dependent Unable to manage medication Finding 285035003 28 CGA IALDs Shopping Independent Shopping independent  Finding 129026007    Assisted Shopping assisted Finding 129073001    Dependent Shopping dependent  Finding 129032002 29 CGA ALDs Bathing Independent Independent bathing Finding 129041007    Assisted Bathing assisted Finding 129040008    Dependent Dependent for bathing Finding 129043005 30 CGA ALDs Dressing Independent Independent with dressing Finding 29035000    Assisted Needs help with dressing  Finding 129039006    Dependent Dependent for dressing  Finding 129065005 31 CGA ALDs Feeding Independent Independent feeding Finding 65224005    Assisted Feeding assisted  Finding 165222009    Dependent Dependent for feeding Finding 129033007 32 CGA ALDs Toileting Independent Independent in toilet  Finding 129062008    Assisted Needs help in toilet Finding 129045003    Dependent Dependent in toilet  Finding 129078005 33 CGA Mobility 5 Times Sit to Stand Attempts Score Freetext No Match     34 CGA Mobility 5 Times Sit to Stand Time Score Freetext No Match     35 CGA Mobility Aid None No Match        Cane Cane, device  Physical Object 87405001  92     Walker Walker Physical Object 705406009    Chair Motorized wheelchair device Physical Object 23366006 36 CGA Mobility Bed Independent Able to move in bed Finding 301681008    Pull No Match        Assisted Assistance with mobility in bed  Difficulty moving in bed Procedure Finding 713138001   301685004      Dependent Unable to move in bed Finding 301682001 37 CGA Mini-Cog Score Freetext Mini-Cog brief cognitive screening test score  Observable Entity 713408000 38 CGA MoCA 7.3 Score Freetext No Match     39 CGA Mood Down Yes Depressed mood Finding 366979004          No Normal Mood Symptoms Finding 134416003 40 CGA Motivation High Increased motivation Finding 86808004    Usual Normal motivation  Finding 64423005    Low Low motivation Finding 26413003 41 CGA Mobility Transfers Independent Independent ability to transfer location Finding 714915006    Stand By No Match        Assisted Able to transfer location with assistance  Finding 719024002    Dependent Dependent ability to transfer location Finding 714916007 42 CGA Mobility Walking Independent Independent walking Finding 165245003    Slow Slow on legs Finding 249898006    Assisted Dependent for walking Finding 427512004    Dependent Dependent for walking Finding 427512004 43 CGA Mobility Walk Outside Independent Able to mobilize outside  Finding 301563003    Assisted Difficulty mobilizing outside  Finding 301568007    Can't Unable to mobilize outside Finding 301564009 44 CGA Nutrition Appetite Within Normal Limits Appetite normal Finding 161825005    Fair No Match        Poor Decrease in appetite Finding 64379006 45 CGA Nutrition Weight Good Normal weight  Finding 43664005    Under Underweight  Finding 248342006    Over Overweight  Finding 238131007    Obese Obese Finding 414915002 46 CGA Pain None No Match        Moderate Moderate pain Finding 50415004    Extreme Severe pain Finding 76948002 47 CGA Sleep Within Normal Limits Good sleep pattern Finding 314939008  93      Disrupted Poor sleep pattern Finding 314938000 48 CGA Speech Within Normal Limits No Match         Impaired Speech Problem Finding 267095009 49 CGA Strength Lower Distal Yes Distal muscle weakness Finding 249942005     No Normal muscle function  Finding 20658008 50 CGA Strength Lower Proximal Yes Proximal muscle weakness Finding 249939004     No Normal muscle function  Finding 20658008 51 CGA Strength Within Normal Limits Normal muscle function  Finding 20658008     Weak Muscle weakness Finding 26544005 52 CGA Strength Upper Distal Yes Distal muscle weakness Finding 249942005     No Normal muscle function  Finding 20658008 53 CGA Strength Upper Proximal Yes Proximal muscle weakness Finding 249939004     No Normal muscle function  Finding 20658008 54 CGA Usual Activities No Problem Able to participate in leisure activities Finding 300738008    Some Problem Difficulty participating in leisure activities  Finding 300742006    Unable Unable to participate in leisure activities Finding 300739000 55 CGA Vision Within Normal Limits Normal vision Finding 45089002     Impaired Abnormal vision Finding 7973008          94   Appendix C  Clinician Consensus #1 The tables below show phase 4 mapping of eCGA data elements to SNOMED CT terms.  Criteria Definition Direct Match The same SNOMED CT code was chosen for the eCGA data element. Partial Match A semantically similar code, or code from the same SNOMED CT hierarchy was chosen, but no direct match.  No Match There was no match between the author and clinician SNOMED CT code selection.    Form Element Description Form Code Response SNOMED Name (Shardae) SNOMED Heirarchy SNOMED Code SNOMED Name (Manpreet) SNOMED Heirarchy  SNOMED Code Original Answer Agreed Terms 2 CGA Balance Within Normal Limits  Balance normal  Finding 298312007 Balance normal Finding 298312007 Direct Match      Impaired  Impairment of balance  Finding 387603000 Poor balance Finding 249985001 Partial Match Agreed to authors code 14 CGA Enough Income Yes Income sufficient to meet needs Finding 224190007 Income sufficient to meet needs Finding 224190007 Direct Match      No Income insufficient to meet needs Finding 224191006 Income insufficient to meet needs Finding 224191006 Direct Match   1 CGA Anxiety No  Normal emotional state   Finding 409068001 No Match     No Match Agreed to No Match    Yes Anxiety  Finding 48694002 Anxiety Finding 48694002 Direct Match   28 CGA IALDs Shopping Independent Shopping independent  Finding 129026007 Shopping independent Finding 129026007 Direct Match     Assisted Shopping assisted Finding 129073001 Shopping assisted Finding 129073001 Direct Match     Dependent Shopping dependent  Finding 129032002 Shopping dependent Finding 129032002 Direct Match    95  26 CGA IALDs Driving Independent Able to drive a car Finding 300634001 Finding related to ability to drive a car Finding 365349005 Partial Match Agree to authors code   Assisted Difficulty driving a car Finding 300638003 No Match     No Match     Dependent Unable to drive a car Finding 300635000 Unable to drive a car Finding 300635000 Direct Match   38 CGA MoCA 7.3 Score Freetext No Match     No Match     Direct Match   55 CGA Vision Within Normal Limits Normal vision Finding 45089002 Normal vision Finding 45089002 Direct Match      Impaired Abnormal vision Finding 7973008 Visual impairment Disorder 397540003 No Match Agreed to authors code 48 CGA Speech Within Normal Limits Able to use verbal communication Finding 288599003 No speech problem Situation 162293002 No Match      Impaired Difficulty using verbal communication Finding 32000005 Speech problem Finding 267095009 No Match Agreed to Clinican Code 22 CGA Hearing Within Normal Limits Hearing normal Finding 162339002 Hearing normal Finding 162339002 Direct Match     Impaired Hearing problem Finding 300228004 Hearing impaired Disorder 15188001 No Match Agreed to authors code 34 CGA Mobility 5 Times Sit to Stand Time Score Freetext No Match     No Match     Direct Match      96  Appendix D  Clinician Consensus #2 The tables below show phase 4 mapping of eCGA data elements to SNOMED CT terms.  Criteria Definition Direct Match The same SNOMED CT code was chosen for the eCGA data element. Partial Match A semantically similar code, or code from the same SNOMED CT hierarchy was chosen, but no direct match.  No Match There was no match between the author and clinician SNOMED CT code selection.   Form Element Description Form Code Response SNOMED Name (Mapper #1) SNOMED Hierarchy SNOMED Code SNOMED Name SNOMED Hierarchy Type SNOMED Code Original Answer Agreed Term CGA IALDs Shopping Independent Shopping independent  Finding 129026007 Shopping Independent Finding 129026007      Assisted Shopping assisted Finding 129073001 Shopping Assisted Finding 129073001      Dependent Shopping dependent  Finding 129032002 Shopping Dependent Finding 129032002     CGA Delirium Yes Delirious Finding 419567006 Altered mental status  Finding 419284004   Agreed to Authors Code   No No Match     State of mind normal  Finding 162716005   Agreed to Authors Code CGA Emotional Other   No Match     No match N/A N/A     CGA Hallucination Yes Hallucinations Finding 7011001 Hallucinations Finding 7011001       No Normal perception Finding 247700009 No match N/A N/A   Agreed to  97  Authors Code CGA ALDs Toileting Independent Independent in toilet  Finding 129062008 Independent in toilet Finding 129062008      Assisted Needs help in toilet Finding 129045003 Needs help in toilet Finding 129045003       Dependent Dependent in toilet  Finding 129078005 Dependent in toilet Finding 129078005     CGA ALDs Dressing Independent Independent with dressing Finding 29035000 Independent with dressing Finding 129035000      Assisted Needs help with dressing  Finding 129039006 Needs help with dressing Finding 129039006       Dependent Dependent for dressing  Finding 129065005 Dependent for dressing Finding  129065005     CGA Sleep Within Normal Limits Good sleep pattern Finding 314939008 Good sleep pattern Finding 314939008       Disrupted Poor sleep pattern Finding 314938000 Poor sleep pattern Finding 314938000     CGA Delusion Yes Delusions Finding 2073000 Delusions Finding 2073000       No No Match     No match N/A N/A     CGA MoCA 7.3 Score Free text No Match     No match N/A N/A     CGA Exercise Frequent Exercises regularly Finding 228448000 Exercises regularly  Finding 228448000      Occasional Gets little exercise Finding 228446001 Gets little exercise Finding 228446001       Not Gets no exercise Finding 228445002 Gets no exercise  Finding  228445002      

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.42591.1-0377757/manifest

Comment

Related Items