key: cord-0948865-rqih7fuq authors: Cesario, Alfredo; Simone, Irene; Paris, Ida; Boldrini, Luca; Orlandi, Armando; Franceschini, Gianluca; Lococo, Filippo; Bria, Emilio; Magno, Stefano; Mulè, Antonino; Santoro, Angela; Damiani, Andrea; Bianchi, Daniele; Picchi, Daniele; Rasi, Guido; Daniele, Gennaro; Fabi, Alessandra; Sergi, Paolo; Tortora, Giampaolo; Masetti, Riccardo; Valentini, Vincenzo; D’Oria, Marika; Scambia, Giovanni title: Development of a Digital Research Assistant for the Management of Patients’ Enrollment in Oncology Clinical Trials within a Research Hospital date: 2021-03-27 journal: J Pers Med DOI: 10.3390/jpm11040244 sha: 343f7ad4fe91cf73635f0c9ade0035ca808ea72d doc_id: 948865 cord_uid: rqih7fuq Clinical trials in cancer treatment are imperative in enhancing patients’ survival and quality of life outcomes. The lack of communication among professionals may produce a non-optimization of patients’ accrual in clinical trials. We developed a specific platform, called “Digital Research Assistant” (DRA), to report real-time every available clinical trial and support clinician. Healthcare professionals involved in breast cancer working group agreed nine minimal fields of interest to preliminarily classify the characteristics of patients’ records (including omic data, such as genomic mutations). A progressive web app (PWA) was developed to implement a cross-platform software that was scalable on several electronic devices to share the patients’ records and clinical trials. A specialist is able to use and populate the platform. An AI algorithm helps in the matchmaking between patient’s data and clinical trial’s inclusion criteria to personalize patient enrollment. At the same time, an easy configuration allows the application of the DRA in different oncology working groups (from breast cancer to lung cancer). The DRA might represent a valid research tool supporting clinicians and scientists, in order to optimize the enrollment of patients in clinical trials. User Experience and Technology The acceptance of participants using the DRA is topic of a future analysis. Cancer care is a complex pathway that is based on a multidisciplinary collaboration among professionals who share the latest evidence and pool their expertise and information through regular communication flows [1] . Multidisciplinary data sharing is an essential approach for tracing patients' pathways, optimizing therapeutic opportunities, and improving healthcare quality. This approach increases evidence-based practice and avoids treating patients outside standardized protocols and recommended guidelines [2, 3] . Clinical trials are imperative for testing novel cancer treatments, advancing the knowledge of care, and determining the best strategies to enhance patients' survival and quality of life outcomes [4, 5] . Nevertheless, the possible lack of communication and real-time synchronization among professionals may produce a fragmentation of services and practices, potentially resulting in the non-optimization of patients' accrual in clinical trials and the limitation of their access to innovative therapeutic solutions [4] [5] [6] . One possible solution can be represented by data sharing approaches, facilitating the enrollment of patients in clinical trials that allow for increasing the chances of recovery, testing novel treatments, and improving knowledge of disease. Less than 5% of the patients are currently enrolled in clinical trials due to logistical issues, a lack of resources, and difficulty in data sharing [4, [7] [8] [9] . Our research hospital has a notable oncological vocation, with nearly 60,000 patients annually accessing our facility with its complex organization in clinical, surgical, and service departments that welcome and manage all of the needs of the cancer patients. Specifically, the Comprehensive Cancer Center coordinates and optimizes all of the cancer related activities, guaranteeing the functionality of specific multidisciplinary working groups and the access to innovative therapies through enrollment in clinical trials or comprehensive interpretation of big data at the institutional and network levels [9] [10] [11] . In order to reduce daily communication inconveniences [12, 13] , a specific platform, called "Digital Research Assistant" (DRA), was developed to report real-time every available clinical trial active within our research hospital and assist clinicians in properly matching patients with the more appropriate studies. The aim of this paper is to show how the DRA was implemented for breast cancer clinical trials to map all of the active studies on this specific disease and encouraging proper patients' enrollment. Its scalability was also evaluated presenting the lung cancer case study. A project manager and two Information and Communication Technology (ICT) professionals started a pilot project with the Breast and Lung Cancer institutional Working Groups, following a user-centered designed approach [14] (Figure 1 ). User-centered designed approach. Context: the program manager identifies who are the primary users of the product, how and why they will use it, what are their needs, and which environment they will use the tool. Requirements: when the context is defined, the program manager identifies the detailed requirements of the product, according to the needs of the user. Design solutions and development: once goals and requirements are settled, the ICT professionals and the project manager design and develop the tool for its usability. Evaluate Product: product designers (in this case, ICT professionals) run usability tests to obtain users' feedback on the product. Healthcare professionals of the involved working groups agreed on nine minimal fields of interest to preliminarily classify the characteristics of patients' records in the platform (Table 1 ) and obtain a quick evaluation of the patients and its possible link to the active and open clinical trials, using breast cancer as a case study. User-centered designed approach. Context: the program manager identifies who are the primary users of the product, how and why they will use it, what are their needs, and which environment they will use the tool. Requirements: when the context is defined, the program manager identifies the detailed requirements of the product, according to the needs of the user. Design solutions and development: once goals and requirements are settled, the ICT professionals and the project manager design and develop the tool for its usability. Evaluate Product: product designers (in this case, ICT professionals) run usability tests to obtain users' feedback on the product. Healthcare professionals of the involved working groups agreed on nine minimal fields of interest to preliminarily classify the characteristics of patients' records in the platform (Table 1 ) and obtain a quick evaluation of the patients and its possible link to the active and open clinical trials, using breast cancer as a case study. Particularly, prognosis and treatment are determined by the stage (TNM classification) of the tumor at the time of diagnosis, but also by the histological/molecular subtype that is obtained with biopsies or in the definitive pathological examination. The DRA was created with the aim to meet several essential clinical and research points: • define an operational app that allows to update data informing all users in real-time; • ensure GDPR-compliant data security; • allow access to both authorized internal and external users (i.e., for multicentric studies); • implement a scalable infrastructure manageable by various specialists (i.e., medical doctors, data managers, research nurses, etc.); and, • develop a matchmaking algorithm between eligible patients and clinical trials. The infrastructure was designed and developed by separating the front-end (i.e., the exposed services) from the back-end (information content) in order to ensure data protection and security ( Figure 2 ). cording to the cellular expression of estrogen (ER: positive or negative) and progesterone (PgR: positive or negative) receptors, epidermal growth factor receptor 2 (HER2: positive or negative) and the proliferation index (Ki67: from 1% to 100%): Particularly, prognosis and treatment are determined by the stage (TNM classification) of the tumor at the time of diagnosis, but also by the histological/molecular subtype that is obtained with biopsies or in the definitive pathological examination. The DRA was created with the aim to meet several essential clinical and research points: • define an operational app that allows to update data informing all users in real-time; • ensure GDPR-compliant data security; • allow access to both authorized internal and external users (i.e., for multicentric studies); • implement a scalable infrastructure manageable by various specialists (i.e., medical doctors, data managers, research nurses, etc.); and, • develop a matchmaking algorithm between eligible patients and clinical trials. The infrastructure was designed and developed by separating the front-end (i.e., the exposed services) from the back-end (information content) in order to ensure data protection and security ( Figure 2) . Infrastructure versatility was then tested using a different case study (lung cancer), to confirm the possibility to easily adapt the platform for indications other than the one used for the first development. Infrastructure versatility was then tested using a different case study (lung cancer), to confirm the possibility to easily adapt the platform for indications other than the one used for the first development. A progressive web app (PWA) was developed to implement a cross-platform software scalable on several electronic devices (i.e., PC, tablet, smartphone). Differently from classic web apps, a PWA that is installed on mobile devices acts as if it was a native app of the device itself, allowing: • to use its functionalities with all the browser through a reference URL, without installation; • to adapt the display according to the screen size of the device; and, • to access its functionalities off-line, guaranteeing data loading by using micro service technologies (APIs). Business logic was developed using the Microsoft, NetCore 3.1 Framework. This software was structured with APIs that make access to data in secure mode with a https protocol scalable and decoupled from the front-end. The app is scalable in terms of the evolution and reutilization of the code, as well as maximization of loading information on the network. The angular open source framework version 9.07 was used to develop the front-end, directly running from the browser after being downloaded from the web server. This choice was taken to have an advance in terms of efficiency, saving the exchange of information between client and server every time that there is a request for action by the user. The SignalR open source framework was used to guarantee a real-time update of data, even when the app is open on a browser. This technology automatically updates information modified by other users, using a two-way channel between the client (browser) and the server (web app). In order to ensure the communication of changes in information to clients, even when they are not connected to the web app, neither it is open, push notifications have been activated using the Google Firebase engine. This open source service allows for sending messages through a web service that transmits notifications to the users of the service. Finally, Microsoft Sql Server 2016 Enterprise Edition was the DBMS used to define the relational model related to this architecture. System access is possible through a hybrid authentication architecture ( Figure 3 ) that allows specialists and healthcare professionals located in various research centers to use the platform: • internal users, through access with personal domain credentials; and, • external users from other research centers (after compiling a standard registration form), through authentication managed internally by the application. A progressive web app (PWA) was developed to implement a cross-platform software scalable on several electronic devices (i.e., PC, tablet, smartphone). Differently from classic web apps, a PWA that is installed on mobile devices acts as if it was a native app of the device itself, allowing: • to use its functionalities with all the browser through a reference URL, without installation; • to adapt the display according to the screen size of the device; and, • to access its functionalities off-line, guaranteeing data loading by using micro service technologies (APIs). Business logic was developed using the Microsoft, NetCore 3.1 Framework. This software was structured with APIs that make access to data in secure mode with a https protocol scalable and decoupled from the front-end. The app is scalable in terms of the evolution and reutilization of the code, as well as maximization of loading information on the network. The angular open source framework version 9.07 was used to develop the front-end, directly running from the browser after being downloaded from the web server. This choice was taken to have an advance in terms of efficiency, saving the exchange of information between client and server every time that there is a request for action by the user. The SignalR open source framework was used to guarantee a real-time update of data, even when the app is open on a browser. This technology automatically updates information modified by other users, using a two-way channel between the client (browser) and the server (web app). In order to ensure the communication of changes in information to clients, even when they are not connected to the web app, neither it is open, push notifications have been activated using the Google Firebase engine. This open source service allows for sending messages through a web service that transmits notifications to the users of the service. Finally, Microsoft Sql Server 2016 Enterprise Edition was the DBMS used to define the relational model related to this architecture. System access is possible through a hybrid authentication architecture ( Figure 3 ) that allows specialists and healthcare professionals located in various research centers to use the platform: • internal users, through access with personal domain credentials; and, • external users from other research centers (after compiling a standard registration form), through authentication managed internally by the application. The app admits three profiles: • System Administrator: enabled to manage configuration features of the app, as described in the "Functionalities" section; • User: enabled to input information about a patient, to enroll and to ask for patients enrollment; and, • Study Manager: enabled to use the same functions of the "User" profile, as well as to manage the creation and modification of trials. "System Administrators" manage the access of internal users (with "User" and "Study Manager" profiles) enabling them to use the app. An HR representative of the research hospital supervises the list of users. From the side menu, the following functionalities are available: Other functionalities include system management and configurations, which are dedicated to "System Administrator" profiles. In particular, the legend includes four entries: • Free (green): the patient can be enrolled in a trial; • Enrolled (red): the patient is currently enrolled in a trial; • Requested Enrollment (yellow): the patient has already been requested for a trial. The "Study Manager" will be able to deny the request or allow the patient in the study; and, Search filters are available for a better user experience. In particular, the legend includes four entries: • Free (green): the patient can be enrolled in a trial; • Enrolled (red): the patient is currently enrolled in a trial; • Requested Enrollment (yellow): the patient has already been requested for a trial. The "Study Manager" will be able to deny the request or allow the patient in the study; and, • Selected for Possible Enrollment: the patient has been selected from a "User" to be evaluated for possible enrollment. Patient enrollment changes according to the logged profile. If the profile is "Study Manager", then the enrollment occurs immediately, otherwise a "User" sends a request to the "Study Manager" of the selected trial, which allows or denies access to the patient in the study. When a patient is accepted, or directly recruited, the "Study Manager" inserts the starting and ending date of the trial. A list of Clinical Trials with their status (i.e., active, suspended, closed) is displayed for all of the profiles ( Figure 5 ). To configure a Clinical Trial, the "User", or the "Study Manager" can enter the information related to the study in which patients can be enrolled ( Figure 6 ). These information are shared with other users, especially those that are interested in the same pathology. In particular, the legend includes four entries: • Free (green): the patient can be enrolled in a trial; • Enrolled (red): the patient is currently enrolled in a trial; • Requested Enrollment (yellow): the patient has already been requested for a trial. The "Study Manager" will be able to deny the request or allow the patient in the study; and, • Selected for Possible Enrollment: the patient has been selected from a "User" to be evaluated for possible enrollment. Patient enrollment changes according to the logged profile. If the profile is "Study Manager", then the enrollment occurs immediately, otherwise a "User" sends a request to the "Study Manager" of the selected trial, which allows or denies access to the patient in the study. When a patient is accepted, or directly recruited, the "Study Manager" inserts the starting and ending date of the trial. A list of Clinical Trials with their status (i.e., active, suspended, closed) is displayed for all of the profiles ( Figure 5 ). To configure a Clinical Trial, the "User", or the "Study Manager" can enter the information related to the study in which patients can be enrolled ( Figure 6 ). These information are shared with other users, especially those that are interested in the same pathology. In particular, the legend includes four entries: • Free (green): the patient can be enrolled in a trial; • Enrolled (red): the patient is currently enrolled in a trial; • Requested Enrollment (yellow): the patient has already been requested for a trial. The "Study Manager" will be able to deny the request or allow the patient in the study; and, • Selected for Possible Enrollment: the patient has been selected from a "User" to be evaluated for possible enrollment. Patient enrollment changes according to the logged profile. If the profile is "Study Manager", then the enrollment occurs immediately, otherwise a "User" sends a request to the "Study Manager" of the selected trial, which allows or denies access to the patient in the study. When a patient is accepted, or directly recruited, the "Study Manager" inserts the starting and ending date of the trial. A list of Clinical Trials with their status (i.e., active, suspended, closed) is displayed for all of the profiles ( Figure 5 ). To configure a Clinical Trial, the "User", or the "Study Manager" can enter the information related to the study in which patients can be enrolled ( Figure 6 ). These information are shared with other users, especially those that are interested in the same pathology. "User" and "Study Manager" profiles can input and modify the Phase of the clinical trial, visible on the selection menu while configuring a trial (Figures 7 and 8) . Matchmaking option. An algorithm then configures the clinical trial. By defining the inclusion-exclusion criteria of a patient in a trial enrollment, these criteria become the rules of the algorithm that allows matchmaking between an eligible patient and a trial. When the "User" inserts a new patient, it is possible to click on the action "assign the patient to a study". This action shows the list of clinical trials for which the patient is eligible. If the patient has characteristics that are coherent with the study, he/she can be enrolled. In particular, omic characteristics (such as genomic mutations) may help achieve a Personalized Medicine approach in oncological clinical trial enrollment. As an enrichment of the services offered by the platform, a connection with the GEmelli NEtwoRk for Analysis and Tests in Oncology and medical Research "Generator"-Real World Data facility is offered to the clinician. Gemelli Generator Real-World Data is a research facility whose aim is the integration of the vast amount of patient data that are available in the Gemelli Data warehouse (about 700 million data items as measured at the end of December 2020). The generator takes care of the integration of these data items, in anonymized form, into specific datamarts, based on appropriate terminological systems, quality-checked and normalized with regard to the information originated from different, heterogeneous data sources, like traditional electronic health records (EHRs), omics data, Patient-Reported Experience Measures (PREMS), and Patient-Reported Outcome Measures (PROMS). Machine Learning and Artificial Intelligence-based methods are at the heart of the Generator infrastructure, allowing for researchers to develop state of the art models, clustering, and decision support systems [16] . After the patient selection phase of the DRA, a simple user interface will give clinicians the opportunity to query the Generator datamarts for the availability of further covariates, referred to the selected patients, that can add more information to what is already present in the DRA core. Full integration between the two systems, at the ICT level, will guarantee an automatic and swift response in a privacy protected environment. In this way, researchers can have a deeper view of the available data and formulate more study hypotheses, based on the large variety of information coming from heterogeneous data sources. Matchmaking option. An algorithm then configures the clinical trial. By defining the inclusion-exclusion criteria of a patient in a trial enrollment, these criteria become the rules of the algorithm that allows matchmaking between an eligible patient and a trial. When the "User" inserts a new patient, it is possible to click on the action "assign the patient to a study". This action shows the list of clinical trials for which the patient is eligible. If the patient has characteristics that are coherent with the study, he/she can be enrolled. In particular, omic characteristics (such as genomic mutations) may help achieve a Personalized Medicine approach in oncological clinical trial enrollment. As an enrichment of the services offered by the platform, a connection with the GEmelli NEtwoRk for Analysis and Tests in Oncology and medical Research "Generator"-Real World Data facility is offered to the clinician. Gemelli Generator Real-World Data is a research facility whose aim is the integration of the vast amount of patient data that are available in the Gemelli Data warehouse (about 700 million data items as measured at the end of December 2020). The generator takes care of the integration of these data items, in anonymized form, into specific datamarts, based on appropriate terminological systems, quality-checked and normalized with regard to the information originated from different, heterogeneous data sources, like traditional electronic health records (EHRs), omics data, Patient-Reported Experience Measures (PREMS), and Patient-Reported Outcome Measures (PROMS). Machine Learning and Artificial Intelligence-based methods are at the heart of the Generator infrastructure, allowing for researchers to develop state of the art models, clustering, and decision support systems [16] . After the patient selection phase of the DRA, a simple user interface will give clinicians the opportunity to query the Generator datamarts for the availability of further covariates, referred to the selected patients, that can add more information to what is already present in the DRA core. Full integration between the two systems, at the ICT level, will guarantee an automatic and swift response in a privacy protected environment. In this way, researchers can have a deeper view of the available data and formulate more study hypotheses, based on the large variety of information coming from heterogeneous data sources. "System Administrators" can insert or modify the characteristics of the setting attributes (that are chosen by the WG) related to the clinical trials (Figures 9 and 10) , while "User" and "Study Manager" profiles can select them directly. "System Administrators" can insert or modify the characteristics of the setting attributes (that are chosen by the WG) related to the clinical trials (Figures 9 and 10) , while "User" and "Study Manager" profiles can select them directly. In the "User Requests" section, all of the requests and their status (accepted, refused, and pending for evaluation) are displayed as well as other users' information requests about a patient or a trial ( Figure 11 ). This section is only accessible to "Study Manager" profiles and allows accepting or declining a request ( Figure 12) . Each row shows a single request with the possibility of examining patient or trial information. This section shows a list of all the clinical trials that the logged profile is responsible for. The "Study Manager" profile also allows editing information about the trials and examining the enrolled patients' full list ( Figure 13 ). In the "User Requests" section, all of the requests and their status (accepted, refused, and pending for evaluation) are displayed as well as other users' information requests about a patient or a trial ( Figure 11 ). In the "User Requests" section, all of the requests and their status (accepted, refused, and pending for evaluation) are displayed as well as other users' information requests about a patient or a trial ( Figure 11 ). This section is only accessible to "Study Manager" profiles and allows accepting or declining a request ( Figure 12) . Each row shows a single request with the possibility of examining patient or trial information. This section shows a list of all the clinical trials that the logged profile is responsible for. The "Study Manager" profile also allows editing information about the trials and examining the enrolled patients' full list ( Figure 13 ). This section is only accessible to "Study Manager" profiles and allows accepting or declining a request ( Figure 12) . Each row shows a single request with the possibility of examining patient or trial information. In the "User Requests" section, all of the requests and their status (accepted, refused, and pending for evaluation) are displayed as well as other users' information requests about a patient or a trial ( Figure 11 ). This section is only accessible to "Study Manager" profiles and allows accepting or declining a request ( Figure 12) . Each row shows a single request with the possibility of examining patient or trial information. This section shows a list of all the clinical trials that the logged profile is responsible for. The "Study Manager" profile also allows editing information about the trials and examining the enrolled patients' full list ( Figure 13 ). This section shows a list of all the clinical trials that the logged profile is responsible for. The "Study Manager" profile also allows editing information about the trials and examining the enrolled patients' full list ( Figure 13 ). This section shows all of the patients preferred by "User" profiles, and preference can be deselected. It is possible to send or accept requests of admission in a trial ( Figure 14) . This section shows all of the patients preferred by "User" profiles, and preference can be deselected. It is possible to send or accept requests of admission in a trial ( Figure 14) . This section shows all of the patients preferred by "User" profiles, and preference can be deselected. It is possible to send or accept requests of admission in a trial ( Figure 14) . In this paper, we described the scale-up customization of the first DRA model on breast cancer to lung cancer thought for a high volume cancer care center. Table 2 shows all the varied characteristics of the patient, except for the "Age" field. In this paper, we described the scale-up customization of the first DRA model on breast cancer to lung cancer thought for a high volume cancer care center. Table 2 shows all the varied characteristics of the patient, except for the "Age" field. • histology: we indicated the most common histological subtypes among lung tumors; • grading has been reported according to 2015 WHO Classification [19] ; • resection margin status, indicating the pathological report of the specimen margin; • genetic test indicates whether the patient has carried out genetic testing for ALK and ROS1 genes, EGFR and KRAS gene mutation and PDL1 expression; and, • therapy type indicates which strategy of care has been adopted. Prognosis and treatment are determined by disease stage (TNM classification), as identified by preliminary diagnostic investigations and histology. Surgery remains the main prognostic factor in the early-stage tumors and the completeness of resection (negative margin status) is a widely recognized factor influencing the long-term results in this setting. Otherwise, in locally advanced and metastatic stages, the molecular characterization represents the main determinant of long-term outcomes, based on the dramatic predictive role of featured biomarkers of activity/efficacy for molecular targeted agents and immunotherapy. Table 3 shows the numerical data from a pilot test of the DRA database. Our solution exploited communication among professionals that are involved in oncological care and cancer research. The DRA we developed allows them to know all of the running clinical trials, guaranteeing all patients the best access to cure and research protocols, reducing the fragmentation of patients' access to the oncological care-path with the multiple therapeutic intersections available in high volume centers (radiotherapy, surgery, and new lines of systemic therapy managed by multiple specialists). This digital tool appeared to be well performing for patients' data sharing within the single institution, but also in setting up networks with other cancer centers facilitating patients' enrollment also for peripheral centers. In fact, in high-patient volume centers, such as our institution, the DRA seems a possible efficient resource to face this issue [20] [21] [22] [23] [24] [25] [26] . At the same time, the platform is easily moldable to the needs of different oncology work groups, as evidenced by the easy customization starting from the model for breast cancer to arrive at that for lung cancer. Taking these considerations together, the platform might represent a valid research tool supporting clinicians and scientists, working in both high-and low-volume centers and the enrollment success rates for each matchmaking run is currently the object of in depth analysis and it will be a topic of future publications. User Experience and Technology Acceptance of participants using the DRA is topic of a second dedicated analysis. The impact of tumor board on cancer care: Evidence from an umbrella review The impact of multidisciplinary team meetings on patient assessment, management and outcomes in oncology settings: A systematic review of the literature Collaboration is the key to cancer research Prospective Evaluation of Cancer Clinical Trial Accrual Patterns: Identifying Potential Barriers to Enrollment Effective stakeholder engagement: Design and implementation of a clinical trial (SWOG S1415CD) to improve cancer care Promoting Scientist-Advocate Collaborations in Cancer Research: Why and How Participation in Cancer Clinical Trials The Role of Clinical Trial Participation in Cancer Research: Barriers, Evidence, and Strategies Standardized data collection to build prediction models in oncology: A prototype for rectal cancer Distributed learning on 20 000+ lung cancer patients-The Personal Health Train Electronic health records, communication, and data sharing: Challenges and opportunities for improving the diagnostic process Improving symptom communication through personal digital assistants: The CHAT (Communicating Health Assisted by Technology) project Use of Mobile Devices to Help Cancer Patients Meet Their Information Needs in Non-Inpatient Settings: Systematic Review The Design of Everyday Things The new TNM-based staging of breast cancer Translational Research in the Era of Precision Medicine: Where We Are and Where We Will Go The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer Toxicity and response criteria of the Eastern Cooperative Oncology Group Lung cancer staging and grading. In World Health Organization Clas-sification of Tumors A Systems Medicine Clinical Platform for Understanding and Managing Non-Communicable Diseases E-HOSPITAL-A Digital Workbench for Hospital Operations and Services Planning Using Information Technology and Algebraic Languages. Stud. Health Technol The Effects of the Digital Platform Support Monitoring and Reminder Technology for Mild Dementia (SMART4MD) for People With Mild Cognitive Impairment and Their Informal Carers: Protocol for a Pilot Randomized Controlled Trial Preventing and Addressing the Stress Reactions of Health Care Workers Caring for Patients With COVID-19: Development of a Digital Platform Digital Peer-Support Platform (7Cups) as an Adjunct Treatment for Women With Postpartum Depression: Feasibility, Acceptability, and Preliminary Efficacy Study Digital platform for improving non-radiologists' and radiologists' interpretation of chest radiographs for suspected tuberculosis-A method for supporting task-shifting in developing countries Initiating Telehealth in a Complex Organization The authors declare no conflict of interest.