Origins in COVID-19 surveillance
Sentinel originated from frustration with the data systems used for contact tracing and surveillance during COVID-19. The systems encountered during that response were not equipped for the combination of changing investigation requirements, large contact-tracing operations and rapidly expanding teams.
Adapting those systems required work on data structures, forms, workflows and reporting while public health operations were already under way. General-purpose customer relationship management systems also required staff to learn interfaces and processes that were not designed around communicable disease investigations.
Operational requirements
Large outbreaks require public health units to increase staffing and investigation capacity quickly. There may be little time to wait for developer changes, train incoming staff in commercial CRM workflows or explain each field change individually.
Questions, case definitions and follow-up requirements can change during the response. Those changes need to reach the people carrying out interviews and public health actions through the system they are using.
Sentinel aims to address two requirements:
- Configuration by the surveillance team. Epidemiologists and administrators with appropriate permissions can configure supported surveillance variables, questionnaires, case definitions and task templates through the application.
- Operational continuity. Interview and task workflows present the configured questions, instructions and calculated information to staff. Routine changes should be reflected in the workflow without requiring each user to interpret the underlying configuration.
This approach reduces reliance on source-code changes for routine surveillance configuration. New integrations or application behaviour may still require development, and changed processes may still require operational instruction.
A foundational product
The initial development phase established a foundational product rapidly. It includes patient and investigation records, contact tracing, outbreaks, laboratory intake, configurable surveys, classification, tasks, review workflows, reporting and access controls.
The data model separates person-level information from investigation-specific information and links investigations to laboratory evidence, contacts, exposures, locations and events. Configuration determines how information is collected, mapped, evaluated and reviewed.
Sustainable development and collaboration
Sentinel is now shifting towards longer-term sustainable development, community participation and collaboration. The next phase places greater emphasis on maintenance, testing, documentation, release processes and prioritisation informed by operational use.
Public health practitioners can contribute disease-specific configuration, review investigation workflows and identify gaps. Testers and implementation teams can evaluate installation, laboratory mappings and data handling. Developers and other contributors can work on defined changes, tests and documentation.
The project remains maintainer-led as this participation develops. Collaboration is intended to broaden the operational perspectives informing the software and support its ongoing maintenance.
Project maintainer
Sentinel is developed by Dr Christian Peut, a specialist general practitioner with experience in communicable disease surveillance and public health operations in South Australia, including COVID-19 surveillance and investigation.
The conference abstract describes the configuration-led approach and its application to changing surveillance and outbreak requirements.
LinkedIn