ABSTRACT
Objectives@#Since protecting patients’ privacy is a major concern in clinical research, there has been a growing need for privacy-preserving data analysis platforms. For this purpose, a federated learning (FL) method based on the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) was implemented, and its feasibility was demonstrated. @*Methods@#We implemented an FL platform on FeederNet, which is a distributed clinical data analysis platform based on the OMOP CDM in Korea. We trained it through an artificial neural network (ANN) using data from patients who received steroid prescriptions or injections, with the aim of predicting the occurrence of side effects depending on the prescribed dose. The ANN was trained using the FL platform with the OMOP CDMs of Kyung Hee University Medical Center (KHMC) and Ajou University Hospital (AUH). @*Results@#The area under the receiver operating characteristic curves (AUROCs) for predicting bone fracture, osteonecrosis, and osteoporosis using only data from each hospital were 0.8426, 0.6920, and 0.7727 for KHMC and 0.7891, 0.7049, and 0.7544 for AUH, respectively. In contrast, when using FL, the corresponding AUROCs were 0.8260, 0.7001, and 0.7928 for KHMC and 0.7912, 0.8076, and 0.7441 for AUH, respectively. In particular, FL led to a 14% improvement in performance for osteonecrosis at AUH. @*Conclusions@#FL can be performed with the OMOP CDM, and FL often shows better performance than using only a single institution's data. Therefore, research using OMOP CDM has been expanded from statistical analysis to machine learning so that researchers can conduct more diverse research.
ABSTRACT
Background@#The most important aspect of a retrospective cohort study is the operational definition (OP) of the disease. We developed a detailed OP for the detection of sodiumglucose cotransporter-2 inhibitors (SGLT2i) related to diabetic ketoacidosis (DKA). The OP was systemically verified and analyzed. @*Methods@#All patients prescribed SGLT2i at four university hospitals were enrolled in this experiment. A DKA diagnostic algorithm was created and distributed to each hospital;subsequently, the number of SGLT2i-related DKAs was confirmed. Then, the algorithm functionality was verified through manual chart reviews by an endocrinologist using the same OP. @*Results@#A total of 8,958 patients were initially prescribed SGLT2i. According to the algorithm, 0.18% (16/8,958) were confirmed to have SGLT2i-related DKA. However, based on manual chart reviews of these 16 cases, there was only one case of SGLT2i-related DKA (positive predictive value = 6.3%). Even after repeatedly narrowing the diagnosis range of the algorithm, the effect of a positive predictive value was insignificant (6.3–10.0%, P > 0.999). @*Conclusion@#Owing to the nature of electronic medical record data, we could not create an algorithm that clearly differentiates SGLT2i-related DKA despite repeated attempts. In all retrospective studies, a portion of the samples should be randomly selected to confirm the accuracy of the OP through chart review. In retrospective cohort studies in which chart review is not possible, it will be difficult to guarantee the reliability of the results.
ABSTRACT
Purpose@#The purpose of the study was to validate the Korean version of Patient-Reported Outcomes Measurement Information System 29 Profile v2.1 (K-PROMIS-29 V2.1) among cancer survivors. @*Materials and Methods@#Participants were recruited from outpatient clinics of the Comprehensive Cancer Center at the Samsung Medical Center in Seoul, South Korea, from September to October 2018. Participants completed a survey questionnaire that included the K-PROMIS-29 V2.1 and the European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire (EORTC QLQ-C30). Principal component analysis and confirmatory factor analysis (CFA) and Pearson’s correlations were used to evaluate the reliability and validity of the K-PROMIS-29 V2.1. @*Results@#The mean age of the study participants was 54.4 years, the mean time since diagnosis was 1.2 (±2.4) years, and 349 (87.3%) completed the entire questionnaire. The Cronbach’s alpha coefficients of the seven domains in the K-PROMIS-29 V2.1 ranged from 0.81 to 0.96, indicating satisfactory internal consistency. In the CFA, the goodness-of-fit indices for the K-PROMIS-29 V2.1 were high (comparative fit index, 0.91 and standardized root-mean-squared residual, 0.06). High to moderate correlations were found between comparable subscales of the K-PROMIS-29 V2.1 and subscales of the EORTC QLQ-C30 (r=0.52-0.73). @*Conclusion@#The K-PROMIS-29 V2.1 is a reliable and valid measure for assessing the health-related quality of life domains in a cancer population, thus supporting their use in studies and oncology trials.
ABSTRACT
Objectives@#The outlook of artificial intelligence for healthcare (AI4H) is promising. However, no studies have yet discussed the issues from the perspective of stakeholders in Korea. This research aimed to identify stakeholders’ requirements for AI4H to accelerate the business and research of AI4H. @*Methods@#We identified research funding trends from the Korean National Science and Technology Knowledge Information Service (NTIS) from 2015 and 2019 using “healthcare AI” and related keywords. Furthermore, we conducted an online survey with members of the Korean Society of Artificial Intelligence in Medicine to identify experts’ opinions regarding the development of AI4H. Finally, expert interviews were conducted with 13 experts in three areas (hospitals, industry, and academia). @*Results@#We found 160 related projects from the NTIS. The major data type was radiology images (59.4%). Dermatology-related diseases received the most funding, followed by pulmonary diseases. Based on the survey responses, radiology images (23.9%) were the most demanding data type. Over half of the solutions were related to diagnosis (33.3%) or prognosis prediction (31%). In the expert interviews, all experts mentioned healthcare data for AI solutions as a major issue. Experts in the industrial field mainly mentioned regulations, practical efficacy evaluation, and data accessibility. @*Conclusions@#We identified technology, regulatory, and data issues for practical AI4H applications from the perspectives of stakeholders in hospitals, industry, and academia in Korea. We found issues and requirements, including regulations, data utilization, reimbursement, and human resource development, that should be addressed to promote further research in AI4H.
ABSTRACT
Personal medical information is an essential resource for research; however, there are laws that regulate its use, and it typically has to be pseudonymized or anonymized. When data are anonymized, the quantity and quality of extractable information decrease significantly.From the perspective of a clinical researcher, a method of achieving pseudonymized data without degrading data quality while also preventing data loss is proposed herein. As the level of pseudonymization varies according to the research purpose, the pseudonymization method applied should be carefully chosen. Therefore, the active participation of clinicians is crucial to transform the data according to the research purpose. This can contribute to data security by simply transforming the data through secondary data processing. Case studies demonstrated that, compared with the initial baseline data, there was a clinically significant difference in the number of datapoints added with the participation of a clinician (from 267,979 to 280,127 points, P < 0.001). Thus, depending on the degree of clinician participation, data anonymization may not affect data quality and quantity, and proper data quality management along with data security are emphasized. Although the pseudonymization level and clinical use of data have a trade-off relationship, it is possible to create pseudonymized data while maintaining the data quality required for a given research purpose. Therefore, rather than relying solely on security guidelines, the active participation of clinicians is important.
ABSTRACT
Objectives@#An increasing emphasis has been placed on the integration of clinical data and patient-generated health data (PGHD), which are generated outside of hospitals. This study explored the possibility of using standard terminologies to represent PGHD for data integration. @*Methods@#We chose the 2020 general health checkup questionnaire of the Korean Health Screening Program as a resource. We divided every component of the questionnaire into entities and values, which were mapped to standard terminologies—Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) version 2020-07-31 and Logical Observation Identifiers Names and Codes (LOINC) version 2.68. @*Results@#Eighty-nine items were derived from the 17 questions of the 2020 health examination questionnaire, of which 76 (85.4%) were mapped to standard terms. Fifty-two items were mapped to SNOMED CT and 24 items were mapped to LOINC. Among the items mapped to SNOMED CT, 35 were mapped to pre-coordinated expressions and 17 to post-coordinated expressions. Forty items had one-to-one relationships, and 17 items had one-to-many relationships. @*Conclusions@#We achieved a high mapping rate (85.4%) by using both SNOMED CT and LOINC. However, we noticed some issues while mapping the Korean general health checkup questionnaire (i.e., lack of explanations, vague questions, and overly narrow concepts). In particular, items combining two or more concepts into a single item were not appropriate for mapping using standard terminologies. Although it is not the case that all items need to be expressed in standard terminology, essential items should be presented in a way suitable for mapping to standard terminology by revising the questionnaire in the future.
ABSTRACT
This study was conducted as a pilot project to evaluate the feasibility of building an integrate dementia platform converging preexisting dementia cohorts from several variable levels. The following four cohorts were used to develop this pilot platform: 1) Clinical Research Center for Dementia of South Korea (CREDOS), 2) Korean Brain Aging Study for Early Diagnosis and Prediction of Alzheimer’s disease (K-BASE), 3) Environmental Pollution-induced Neurological Effects (EPINEF) study, and 4) a prospective registry in Dementia Platform Korea project (DPKR). A total of 29916 patients were included in the platform with 348 integrated variables. Among participants, 13.9%, 31.5%, and 44.2% of patients had normal cognition, mild cognitive impairment, and dementia, respectively. The mean age was 72.4 years. Females accounted for 65.7% of all patients. Those with college or higher education and those without problems in reading or writing accounted for 12.3% and 46.8%, respectively. Marital status, cohabitation, family history of Parkinson’s disease, smoking and drinking status, physical activity, sleep status, and nutrition status had rates of missing information of 50% or more. Although individual cohorts were of the same domain and of high quality, we found there were several barriers to integrating individual cohorts, including variability in study variables and measurements. Although many researchers are trying to combine pre-existing cohorts, the process of integrating past data has not been easy. Therefore, it is necessary to establish a protocol with considerations for data integration at the cohort establishment stage.
ABSTRACT
This study was conducted as a pilot project to evaluate the feasibility of building an integrate dementia platform converging preexisting dementia cohorts from several variable levels. The following four cohorts were used to develop this pilot platform: 1) Clinical Research Center for Dementia of South Korea (CREDOS), 2) Korean Brain Aging Study for Early Diagnosis and Prediction of Alzheimer’s disease (K-BASE), 3) Environmental Pollution-induced Neurological Effects (EPINEF) study, and 4) a prospective registry in Dementia Platform Korea project (DPKR). A total of 29916 patients were included in the platform with 348 integrated variables. Among participants, 13.9%, 31.5%, and 44.2% of patients had normal cognition, mild cognitive impairment, and dementia, respectively. The mean age was 72.4 years. Females accounted for 65.7% of all patients. Those with college or higher education and those without problems in reading or writing accounted for 12.3% and 46.8%, respectively. Marital status, cohabitation, family history of Parkinson’s disease, smoking and drinking status, physical activity, sleep status, and nutrition status had rates of missing information of 50% or more. Although individual cohorts were of the same domain and of high quality, we found there were several barriers to integrating individual cohorts, including variability in study variables and measurements. Although many researchers are trying to combine pre-existing cohorts, the process of integrating past data has not been easy. Therefore, it is necessary to establish a protocol with considerations for data integration at the cohort establishment stage.
ABSTRACT
Subject(s)
Humans , Diagnosis , Focus Groups , Health Occupations , Information Dissemination , Korea , Methods , Precision Medicine , Qualitative Research , Surveys and QuestionnairesABSTRACT
In recent years, artificial intelligence (AI) technologies have greatly advanced and become a reality in many areas of our daily lives. In the health care field, numerous efforts are being made to implement the AI technology for practical medical treatments. With the rapid developments in machine learning algorithms and improvements in hardware performances, the AI technology is expected to play an important role in effectively analyzing and utilizing extensive amounts of health and medical data. However, the AI technology has various unique characteristics that are different from the existing health care technologies. Subsequently, there are a number of areas that need to be supplemented within the current health care system for the AI to be utilized more effectively and frequently in health care. In addition, the number of medical practitioners and public that accept AI in the health care is still low;moreover, there are various concerns regarding the safety and reliability of AI technologyimplementations. Therefore, this paper aims to introduce the current research and application status of AI technology in health care and discuss the issues that need to be resolved.
ABSTRACT
Recently, digital health has gained the attention of physicians, patients, and healthcare industries. Digital health, a broad umbrella term, can be defined as an emerging health area that uses brand new digital or medical technologies involving genomics, big data, wearables, mobile applications, and artificial intelligence. Digital health has been highlighted as a way of realizing precision medicine, and in addition is expected to become synonymous with health itself with the rapid digitization of all health-related data. In this article, we first define digital health by reviewing the diverse range of definitions among academia and government agencies. Based on these definitions, we then review the current status of digital health, mainly in Korea, suggest points that are missing from the discussion or ought to be added, and provide future directions of digital health in clinical practice by pointing out certain key points.
Subject(s)
Humans , Artificial Intelligence , Genomics , Government Agencies , Government Regulation , Health Care Sector , Korea , Mobile Applications , Precision Medicine , TelemedicineABSTRACT
OBJECTIVES: For earlier detection of infectious disease outbreaks, a digital syndromic surveillance system based on search queries or social media should be utilized. By using real-time data sources, a digital syndromic surveillance system can overcome the limitation of time-delay in traditional surveillance systems. Here, we introduce an approach to develop such a digital surveillance system. METHODS: We first explain how the statistics data of infectious diseases, such as influenza and Middle East Respiratory Syndrome (MERS) in Korea, can be collected for reference data. Then we also explain how search engine queries can be retrieved from Google Trends. Finally, we describe the implementation of the prediction model using lagged correlation, which can be calculated by the statistical packages, i.e., SPSS (Statistical Package for the Social Sciences). RESULTS: Lag correlation analyses demonstrated that search engine data/Twitter have a significant temporal relationship with influenza and MERS data. Therefore, the proposed digital surveillance system can be used to predict infectious disease outbreaks earlier. CONCLUSIONS: This prediction method could be the core engine for implementing a (near-) real-time digital surveillance system. A digital surveillance system that uses Internet resources has enormous potential to monitor disease outbreaks in the early phase.
Subject(s)
Communicable Diseases , Coronavirus Infections , Disease Outbreaks , Influenza, Human , Information Storage and Retrieval , Internet , Korea , Methods , Search Engine , Social MediaABSTRACT
Recent rapid advances in artificial intelligence (AI), especially in deep learning methods, have produced meaningful results in many areas. However, to achieve meaningful results for healthcare through AI, it is important to understand the meaning and characteristics of data in that area. For medical AI, a simple approach that accumulates massive amounts of data based on existing big data concepts cannot provide meaningful results in the healthcare field. We need well-curated data as opposed to a simple aggregation of data. The purpose of this study is to present the types and characteristics of healthcare data and future directions for the successful combination of AI and medical care.
Subject(s)
Artificial Intelligence , Delivery of Health Care , Korea , Learning , Machine LearningABSTRACT
No abstract available.
Subject(s)
Child , Humans , Data Collection , Delivery of Health Care , Online Systems , Text MessagingABSTRACT
OBJECTIVES: To evaluate the mobile health applications (apps) developed by a single tertiary hospital in Korea with a particular focus on quality and patient safety. METHODS: Twenty-three mobile health apps developed by Asan Medical Center were selected for analysis after exclusion of the apps without any relationship with healthcare or clinical workflow, the apps for individual usage, and the mobile Web apps. Two clinical informaticians independently evaluated the apps with respect to the six aims for quality improvement suggested by the United States Institute of Medicine. All discrepancies were resolved after discussion by the two reviewers. The six aims observed in the apps were reviewed and compared by target users. RESULTS: Eleven apps targeted patients, the other 12 were designed for healthcare providers. Among the apps for patients, one app also had functions for healthcare providers. 'My cancer diary' and 'My chart in my hand' apps matched all the six aims. Of the six aims, Timeliness was the most frequently observed (20 apps), and Equity was the least observed (6 apps). Timeliness (10/11 vs. 10/12) and Patient safety (10/11 vs. 9/12) were frequently observed in both groups. In the apps for patients, Patient-centeredness (10/11 vs. 2/12) and Equity (6/11 vs. 0/12) were more frequent but Efficiency (5/11 vs. 10/12) was less frequent. CONCLUSIONS: Most of the six aims were observed in the apps, but the extent of coverage varied. Further studies, evaluating the extent to which they improve quality are needed.
Subject(s)
Humans , Delivery of Health Care , Health Personnel , Korea , Patient Safety , Patient-Centered Care , Quality Improvement , Telemedicine , Tertiary Care Centers , United StatesABSTRACT
De-identification of personal health information is essential in order not to require written patient informed consent. Previous de-identification methods were proposed using natural language processing technology in order to remove the identifiers in clinical narrative text, although these methods only focused on narrative text written in English. In this study, we propose a regular expression-based de-identification method used to address bilingual clinical records written in Korean and English. To develop and validate regular expression rules, we obtained training and validation datasets composed of 6,039 clinical notes of 20 types and 5,000 notes of 33 types, respectively. Fifteen regular expression rules were constructed using the development dataset and those rules achieved 99.87% precision and 96.25% recall for the validation dataset. Our de-identification method successfully removed the identifiers in diverse types of bilingual clinical narrative texts. This method will thus assist physicians to more easily perform retrospective research.
Subject(s)
Humans , Algorithms , Data Anonymization , Electronic Health Records , Health Records, Personal , Multilingualism , Natural Language Processing , Research DesignABSTRACT
OBJECTIVES: Due to the unique characteristics of clinical data, clinical data warehouses (CDWs) have not been successful so far. Specifically, the use of CDWs for biomedical research has been relatively unsuccessful thus far. The characteristics necessary for the successful implementation and operation of a CDW for biomedical research have not clearly defined yet. METHODS: Three examples of CDWs were reviewed: a multipurpose CDW in a hospital, a CDW for independent multi-institutional research, and a CDW for research use in an institution. After reviewing the three CDW examples, we propose some key characteristics needed in a CDW for biomedical research. RESULTS: A CDW for research should include an honest broker system and an Institutional Review Board approval interface to comply with governmental regulations. It should also include a simple query interface, an anonymized data review tool, and a data extraction tool. Also, it should be a biomedical research platform for data repository use as well as data analysis. CONCLUSIONS: The proposed characteristics desired in a CDW may have limited transfer value to organizations in other countries. However, these analysis results are still valid in Korea, and we have developed clinical research data warehouse based on these desiderata.
Subject(s)
Anonyms and Pseudonyms , Ethics Committees, Research , Ethics, Research , Information Storage and Retrieval , Korea , Privacy , Research Design , Social Control, Formal , Statistics as TopicABSTRACT
OBJECTIVES: The Korean government has enacted two laws, namely, the Personal Information Protection Act and the Bioethics and Safety Act to prevent the unauthorized use of medical information. To protect patients' privacy by complying with governmental regulations and improve the convenience of research, Asan Medical Center has been developing a de-identification system for biomedical research. METHODS: We reviewed Korean regulations to define the scope of the de-identification methods and well-known previous biomedical research platforms to extract the functionalities of the systems. Based on these review results, we implemented necessary programs based on the Asan Medical Center Information System framework which was built using the Microsoft. NET Framework and C#. RESULTS: The developed de-identification system comprises three main components: a de-identification tool, a search tool, and a chart review tool. The de-identification tool can substitute a randomly assigned research ID for a hospital patient ID, remove the identifiers in the structured format, and mask them in the unstructured format, i.e., texts. This tool achieved 98.14% precision and 97.39% recall for 6,520 clinical notes. The search tool can find the number of patients which satisfies given search criteria. The chart review tool can provide de-identified patient's clinical data for review purposes. CONCLUSIONS: We found that a clinical data warehouse was essential for successful implementation of the de-identification system, and this system should be tightly linked to an electronic Institutional Review Board system for easy operation of honest brokers. Additionally, we found that a secure cloud environment could be adopted to protect patients' privacy more thoroughly.
Subject(s)
Humans , Access to Information , Bioethics , Computer Security , Electronics , Electrons , Ethics Committees, Research , Ethics, Research , Information Systems , Jurisprudence , Masks , Privacy , Research Design , Social Control, Formal , Tertiary Care CentersABSTRACT
We have noticed an inadvertent error in our article. In Figure 1, an abbreviation is misspelled.