Paediatric Rapid Sepsis Trigger (PRST) Tool (PRST)

Smart Triage: Triage and Management of Sepsis in Children Using the Point-of-care Paediatric Rapid Sepsis Trigger (PRST) Tool

Sepsis is the leading cause of death and disability in children, every hour of delay in treatment is associated with greater organ damage and ultimately death. The challenges, especially in poor countries, are the delays in diagnosis and the inability to identify children in urgent need of treatment.To circumvent these challenges, we propose the development and clinical evaluation of a trigger tool that will reduce the time to diagnosis and prompt the timely initiation of life-saving treatment. The key innovations are 1) a data-driven approach to rapid diagnosis of sepsis severity and 2) a low- cost digital tagging system to track the time to treatment. The tool will require minimal cost, clinical expertise and training or time to use. The tool will identify high risk children and reduce time to treatment. The digital platform (mobile device and dashboard) will create a low-cost, highly scalable solution for children with sepsis.

This is a pre-post intervention study involving pediatric patients presenting to the study hospitals in seek of medical care for an acute illness. The study involves three phases: (I) Baseline Period, (II) Interphase Period, (III) Intervention Period. Phase I (3-6 months) will be a prospective observational cohort at Mbagathi County Hospital in Nairobi, Kenya, and Jinja Regional Referral Hospital in Jinja, Uganda. During this period, there will be no changes to healthcare delivery procedures in the study hospitals. Triage will continue to be performed by hospital staff using Emergency Triage and Treatment (ETAT) guidelines, the system that is currently in effect at the study hospitals. Data collection will be undertaken in the triage waiting area. While participants are waiting for their turn to be seen by the hospital triage nurses, our trained study nurses will collect data on a pre-selected list of predictor variables. These data will be used to develop a clinical prediction model based on the need for hospital admission. Control Site (Phase I, 12 months): Kiambu County Referral Hospital in Nairobi, Kenya will serve as the control site and no intervention will be implemented. At this site, Phase I will commence for a period of 12 months. There will be no Phase II or Phase III. Phase II (1-3 months) will involve technology development, usability testing, and training. Phase IIa: Technology Development. A risk prediction model will be derived using the data collected in Phase I and implemented in a Digital Triaging Platform, along with a digitized version of the ETAT+ guidelines. The Digital Triaging Platform, including vital sign measurement devices (PhoneOx and RRate and the mobile application and clinical dashboard has already been developed and evaluated. Once the digital triage tool has been developed, it will be evaluated in potential users using simulated patient scenarios and a 'Think Aloud' method. Phase IIb: Usability Testing and Training. The digital triage tool will be evaluated for ease of interface navigation, functionality and basic workflow. A sample of 15 health workers in the study hospitals to represent the primary user groups will be selected for participation in the 60-minute-long usability testing initiatives. Participants will be recruited through word of mouth as there is a very small cadre of potential participants. The objective of the training is to (1) ensure healthcare workers understand how to correctly collect and interpret patient information, and (2) to obtain feedback on the digitization of the tool. Training will use a framework that meets key international norms for testing digital tools, including, the think-aloud method and a questionnaire. Each training session will be conducted by a moderator and observer. During the evaluation, the observer will be seated next to the participant and will record user interaction with each interface, comments, errors, and duration of each task. Participants will be given 3-5 patient scenarios which will list hypothetical information to be entered into the app. This information will be designed to represent routine data collected during triage examination at the study hospitals. The moderator will provide the fictional charts to participants and instruct them throughout the tasks. During the simulated patient scenarios, participants will be asked to think aloud, in order to assess their thought process as they used the app. Participants will be specifically instructed to comment on the layout of the app screen, the dialogue on each interface, the order of tasks, and any additional observations or opinions. After learning the basics of the digital platform, the participants will be read the think aloud instructions and asked to perform the list of tasks and answer questions. The observer will complete a checklist to ensure that all tasks were completed, questions will be asked to evaluate task comprehension, and notes will be taken about whether help was needed in completing each task. At the end of the training session, participants will complete a triage tool training questionnaire to provide an understanding of the practical benefits and drawbacks of incorporating the digital triage tool into a clinical context. The questionnaire will utilize open ended questions and comment responses. from this evaluation. Responses from the survey will be anonymous. The data generated from the training phase is fictitious and will not be linked to any individual subject. Transcriptions and Think Aloud observations will be analyzed using the Framework Method to assess attitudes of health workers. Responses will be transcribed and coded for the identification, examination and interpretation of emerging themes and patterns. Results from the analysis, feedback from the questionnaires, and comments on the observer checklists will be used to generate a report with suggested improvements to be shared with the quality improvement implementation team prior to Phase III. Phase III (3-6 months) will be an interventional period involving routine use of the digital triage tool by the hospital triage nurses at Mbagathi County Hospital in Nairobi, Kenya, and Jinja Regional Referral Hospital in Jinja, Uganda. The digital triage tool will not replace triage policies already in place at the study hospitals, but rather it will supplement and strengthen existing triage systems. As done in Phase I, study nurses will collect data on the pre-selected list of predictor variables in the triage waiting area while participants are waiting to be seen by the hospital triage nurses (who will be using the digital triage tool). Continued collection of predictor variables will allow comparison of participant characteristics in the pre-intervention cohort and the post-intervention cohort.

This is a pre-post intervention study involving pediatric outpatients at Jinja Hospital in Jinja, Uganda, and Mbagathi Hospital in Nairobi, Kenya. The study has three phases: (I) Baseline Period: data is collected on key predictors and outcomes before implementation of the digital triage tool, (II) Interphase Period: model/technology development and usability testing, (III) Intervention Period: data is collected on key predictors and outcomes after implementation of the digital triage tool. Patients who present to the study hospitals in seek of medical care for an acute illness during Phase I will be part of the baseline cohort. Patients who present to the study hospitals in seek of medical care for an acute illness during Phase III will be part of the interventional cohort. A third hospital site, Kiambu Hospital in Nairobi Kenya, will serve as a control hospital, and will only participate in Phase I for the entire duration of the study.

Hospital nurses will triage participants using the Emergency Triage and Treatment (ETAT) guidelines, the triage policy that is currently in place at the study hospital sites.

The digital platform consists of a mobile application integrating a pulse oximetry sensor attached to this device, with embedded smart algorithms that predict a critically ill state, or level of risk (below) in a child presenting at the hospital. The platform also includes an interactive dashboard located in strategic locations (e.g., laboratory, consultation rooms), which connects to the mobile application through a secure local network and displays the triage data to provide real-time monitoring for the physicians who manage the patients.

These are the triage guidelines currently applied in the study hospital sites. This will be the comparator group.

Model Development: Hospital admission (within 5 days of assessment) status determined from hospital records and information system, and a follow up call 7 days post discharge (to ascertain outcome). This will inform development of a clinical prediction model based on need for hospital admission.

Effectiveness Evaluation of Digital Triage Tool: Time to administration of an appropriate sepsis bundle of care, which includes at least one of antibiotics, intravenous fluids, or oxygen as appropriate for age and clinical syndrome (treatment determined and administered by hospital staff). This will allow us to quantify the change from baseline in the proportion of children in each triage category (emergency, priority, queue) receiving treatment within 60 minutes of arrival at hospital.

Length of hospitalization determined from hospital records, and a follow up call 7 days post-discharge.

Inclusion Criteria: All paediatric outpatients presenting to the study hospitals for medical treatment. The lower age limit will include children aged from 0 days, and the upper age limit will be in accordance to respective hospitals' practice for paediatric admissions (this may be 12, 15 or 19 years). Informed parental/guardian consent provided. Assent from children older than 13 years in addition to parental/guardian consent provided. Exclusion Criteria: -Patients presenting to the outpatient department for elective cases (e.g. elective surgery or change of dressing) or for clinical review appointment.

After the study period, a de-identified copy of the data will be prepared for deposition in a repository with open access with proper governance mechanisms. We will make every effort to prevent re-identification of subjects by coding data that has the potential of being identifiable. The de-identified study data will be made publicly available using the Harvard Dataverse (https://dataverse.harvard.edu/), which is the data repository for KWTRP, and a URL will be made accessible. To enhance visibility, sharing and collating datasets with other collaborating sites for increased usability/re-use, de-identified will also be shared availed to reputable data hosting service such as the INDEPTH Data Repository (http://www.indepth-ishare.org/index.php/home), or through the newly established Pediatric Sepsis CoLab (sponsored by the World Federation of Pediatric Critical and Intensive care Societies).

Sharing and access will be managed and subject to institutional agreements (Kenya Medical Research Institute, The University of British Columbia, Makerere University School of Public Health) that will set terms for how requests and access will be managed. We will ensure that a rigorous data governance structure is used by the data hosting service. The distribution will only occur with agreement from Principal Investigators and the investigators at all of the study sites.

Li ECK, Grays S, Tagoola A, Komugisha C, Nabweteme AM, Ansermino JM, Mitton C, Kissoon N, Khowaja AR. Cost-effectiveness analysis protocol of the Smart Triage program: A point-of-care digital triage platform for pediatric sepsis in Eastern Uganda. PLoS One. 2021 Nov 17;16(11):e0260044. doi: 10.1371/journal.pone.0260044. eCollection 2021.

Mawji A, Li E, Komugisha C, Akech S, Dunsmuir D, Wiens MO, Kissoon N, Kenya-Mugisha N, Tagoola A, Kimutai D, Bone JN, Dumont G, Ansermino JM. Smart triage: triage and management of sepsis in children using the point-of-care Pediatric Rapid Sepsis Trigger (PRST) tool. BMC Health Serv Res. 2020 Jun 3;20(1):493. doi: 10.1186/s12913-020-05344-w.